[ { "id": 28841, "name": "Baud", "slug": "baud", "former_names": [ "Neuralyze", "Cerelyze", "Zener", "Front", "Zener", "SuperCraft", "Ponder", "SuperCraft", "Baud Labs" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/2ea89d0878c367c79b3a58c19eb1323a648e3f38.png", "website": "https://baudlabs.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Novel AI hardware built from first principles to let companies train, fine-tune, and deploy frontier models with orders of magnitude better performance than incumbents", "one_liner": "AI chips for ultra-fast training and inference", "team_size": 3, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1688534176, "tags": [ "Hard Tech", "Semiconductors", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/baud", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/baud.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/baud-labs-ai/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/2ea89d0878c367c79b3a58c19eb1323a648e3f38.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 1348413, "full_name": "Sarang Zambare", "title": "Founder/CEO", "founder_bio": "Multi-time founder with 4 patents and 7+ years in deep learning and AI hardware. ML lead for Peloton Guide from concept to 100,000+ devices shipped. Previously founding ML engineer at Peloton and Caper (acquired by Instacart). Also a certified pilot.", "is_active": true, "linkedin_url": "https://linkedin.com/in/sarang-zambare", "x_url": "https://twitter.com/sarangz11", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d9ae5ac000c0c036f4b5022a8dd94a1b90a9e2a9.jpg" }, { "user_id": 3805085, "full_name": "Eric Taylor", "title": "Founder & Chief Hardware Architect", "founder_bio": "Experienced ASIC Architect with over a decade of industry experience, including four tape-outs, two major IP releases, two patents, and a publication. Prior roles at NVIDIA, Freescale/NXP, Arteris IP, and Enfabrica.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/eric-taylor-995b923b/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/54e1b3deed285b11303c26e69ac744fbf80d5f4c.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 104577, "title": "Baud - AI Chips for Accelerated Training and Inference", "tagline": "Novel AI hardware and developer platform, co-designed from first principles to let businesses create frontier models with orders of magnitude better performance than incumbents", "body": "**TL;DR**\n\nBaud is building a new chip for training and inference of large AI models\n\nWe have developed a new arithmetic representation of neural networks that does not contain multiplications. Our silicon is architected specifically for this representation and doing this lets us train and serve AI faster and more efficiently than incumbents.\n\n![uploaded image](/media/?type=post&id=104577&key=user_uploads/1348413/43bfbd38-2274-4963-b225-3714d18f7c65)\n\n\\\n\\\n**Our Team** \n\n* **Sarang Zambare**, multi-time founder with 4 patents and 7+ years in deep learning and AI hardware. ML lead for Peloton Guide from concept to 100,000+ devices shipped and founding engineer at Caper (acquired by Instacart) building edge deep learning models.\n* **Eric Taylor**, ex-NVIDIA chip design veteran with a decade of experience including 4 tape-outs on leading process nodes, two major IP releases, multiple patents and publications, prior roles at NVIDIA, Freescale/NXP, Arteris IP, and Enfabrica.\n\n\\\n\\\n**The Problem**\n\nWe are living in exciting times where humanity has figured out a way to turn sand and sunlight into machines of scientific discovery.\n\nBut creating and serving these machines aka frontier AI models, is extremely capital intensive.\n\nNVIDIA’s publicly documented training recipe of the Nemotron model took **6144 H100 GPUs**, going on for around 3 months, drawing around **4.3 megawatts** of electricity.\n\n![uploaded image](/media/?type=post&id=104577&key=user_uploads/1348413/eddfed11-11fd-41b3-bbc2-8bb950078002)\n\nThat's **$44M in rentals or $245M in capex** plus an ungodly number of developer hours standing up distributed training. Models like Claude Opus 4.6 surprised the world, and that's just the beginning. But at these numbers, **only a handful of teams on Earth control the frontier in AI** and have a shot at a new opus-level moment in other domains like robotics, world models, drug discovery, material science, physics, simulation, the list goes on.\\\n\\\nCurrent attempts at solving this such as Cerebras, Groq, Etched, D-Matrix and 50 other chip startups are only attacking the silicon side and leaving the fundamental math side of the problem untouched, because they are focused on inference and hence have to run existing model weights.\n\nWhen we looked deeper, we found that if you let go of the constraint of running existing model weights, a radically different silicon architecture becomes possible, with an order of magnitude better performance.\n\nWe think we're just scratching the surface of what's possible when you co-design the hardware with the model architecture.\n\nAnd that’s what we are betting the farm on.\n\n**Solution**\n\nWe have developed a new way to represent neural network architectures that eliminates multiplications from both the forward and backward passes, and also compresses the weights by more than 10x without any loss in intelligence. The catch is that you need to train the model in this representation, or start with a base model that is also in this representation.\n\nBoth during training and inference, our chip does not need multiplier circuits.\n\nThis makes our individual cores drastically smaller and simpler than say GPU’s tensor cores or the PEs in TPU’s systolic arrays.\n\nWhich means we can pack way more compute and SRAM on the die and move way more weights for the same memory bandwidth.\n\nThis makes our chip faster, less power hungry and simpler to build.\\\n\\\nWe provide a developer platform based on our hardware that makes creation of frontier AI frictionless and economical to serve, so businesses of all sizes can start owning their intelligence.\\\n\\\n**We are live already**\n\n![uploaded image](/media/?type=post&id=104577&key=user_uploads/1348413/fb27985c-5941-4f2b-aac1-a81378047b7c)\n\nOur first chip is validated on the Global Foundries 12nm process and on schedule for tape-out by the end of this year.\\\n\\\nOur training and inference service is already live on a cluster of FPGAs running the chip architecture in emulation.\\\n\\\nUsing FPGA emulation, we've developed a compiler and a distributed training stack that converts any PyTorch exportable model to our format with bit exact results in most cases, including popular opensource architectures such as GLM, Qwen, Gemma, DeepSeek, Flux, Wan and others.\\\n\\\nYou can checkout a small demo of our chip running in emulation on a single FPGA here: \n\n**Our partner program**\\\n\\\nBaud is now working with design partners who can test our systems and help smooth out the developer experience, in exchange for reserved capacity on our first cluster for training and inference.\\\n\\\nCapacity is limited. Contact us if you are training something new or are tired of paying through the nose for intelligence that you do not own.\\\n\\\n\n\n![uploaded image](/media/?type=post&id=104577&key=user_uploads/1348413/7e141e34-4913-4656-b502-b1b7201eeb73)\n\n", "url": "https://www.ycombinator.com/launches/RCj-baud-ai-chips-for-accelerated-training-and-inference", "votes": 98, "created_at": "2026-07-06T15:52:27.467Z" } ] }, { "id": 30653, "name": "Async", "slug": "withasync", "former_names": [ "Skope" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/1c746a958222a919c90dfff3a0cd3a949fa81bcc.png", "website": "https://withasync.com", "all_locations": "New York City, NY, USA", "long_description": "Async builds AI agents that run small businesses. Law firms, healthcare clinics and real estate companies hire our AI agents to automate complex work that's never been economical to staff with humans, letting us capture labor spend that didn't previously exist.", "one_liner": "AI agents that run small businesses", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1750879126, "tags": [ "B2B", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/withasync", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/withasync.json", "year_founded": 2025, "city": "New York City", "country": "US", "linkedin_url": "https://www.linkedin.com/company/useskope/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/1c746a958222a919c90dfff3a0cd3a949fa81bcc.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/0a8c3ad6fd2022180a90d5d995561b3d2fe2297f.png" ], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 2087375, "full_name": "Ben Smith", "title": "Founder/CEO", "founder_bio": "Co-Founder of Async. Used to sell shoes, trading cards and collectibles. Dropped out of college at 20 to build software. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ben-smith-b15b12293/", "x_url": "https://x.com/imbensmith", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d9ce70059aceb2cd015e416ccb6f9b0525ad8c0a.jpg" }, { "user_id": 1593011, "full_name": "Connor Park", "title": "Founder/CTO", "founder_bio": "Previously studied Computer Science at the University of Michigan, Ann Arbor. Building since 7 years old. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/connor-park24/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/62122d6c94fbef6ba7cb8070138df793a1d155da.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109673, "title": "Async - Specialized AI agents, that automate costly work for small businesses.", "tagline": "AI agents that change how small businesses can run", "body": "[https://drive.google.com/file/d/1pDRqFr9HS9Tlou6LoJb8I9sFyKTJpDbv](https://drive.google.com/file/d/1pDRqFr9HS9Tlou6LoJb8I9sFyKTJpDbv/view?usp=sharing)\n\\\nTLDR: Async builds AI agents that change how real economy small businesses can run. Law firms, healthcare clinics, real estate companies, and childcare facilities hire our AI agents to automate complex work that's never been economical to staff with humans. Every day, our agents complete work from scheduling clinical therapy appointments, to drafting long legal documents, to entering data into legacy practice management systems. We've automated thousands of hours worth of work for small businesses around the country.\\\n\\\n**The Problem:**\\\nSmall businesses lack the ability to effectively adopt AI.\n\n* Effective adoption requires rebuilding operating procedures to be agent first, instead of human first. SMB owners generally lack the time and knowledge to do this.\n* Out-of-the-box solutions fail to make an outsized impact within complex operations.\n* These businesses cannot grow without scaling headcount.\n\n\\\n**The Solution:**\\\nWe rebuild workflows and deploy AI agents to automate them end-to-end.\n\n* We build a configurable platform that lets us quickly deploy specialized AI agents inside of their business.\n* These agents work within our platform and on top of legacy systems.\n* The work our agents handle for customers spans everything from drafting real-estate closing documents, to coordinating scheduling for ABA therapy appointments, to handling data entry into legacy legal software systems.\n\n\\\n**Asks:**\n\n* Introductions to small businesses (especially within law, healthcare, real estate, and childcare) would be greatly appreciated.", "url": "https://www.ycombinator.com/launches/SWv-async-specialized-ai-agents-that-automate-costly-work-for-small-businesses", "votes": 8, "created_at": "2026-08-10T17:30:00.428Z" } ] }, { "id": 30672, "name": "Induction Labs", "slug": "induction-labs", "former_names": [ "Next Action" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/0c2f926be6434de6b8c75a3a873f2b79fe0bdf59.png", "website": "https://inductionlabs.com", "all_locations": "San Francisco, CA, USA", "long_description": "We are scaling foundation models that are curious about the world.\r\n\r\nWe’ve seen general intelligence appear exactly once, and its hallmark has been curiosity: a dissatisfaction with what we already know about the world, an ability to learn from world experience, and a disposition to share knowledge with other individuals. Across generations, these human traits have built modern science, technology, and culture.\r\n\r\nWe think superintelligence will come from models that work the same way: learning from their own observation, updating as they go, growing what they know. These models will start from what humanity knows and discover past us.\r\n\r\nWe believe that scaling curious intelligence is a defining problem of our time. If solved, we will live to see a new era of technological advances, scientific understanding, and human flourishing.", "one_liner": "Building intellectually curious AI", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1755632182, "tags": [ "Deep Learning", "Hard Tech", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/induction-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/induction-labs.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/induction-labs", "x_url": "https://x.com/induction_labs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/0c2f926be6434de6b8c75a3a873f2b79fe0bdf59.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harj Taggar", "url": "https://www.ycombinator.com/people/harj-taggar" }, "founders": [ { "user_id": 2757445, "full_name": "Jonathan Li", "title": "Founder/CEO", "founder_bio": "Building foundation models of the future @ Induction Labs", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jonatli/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/90c27a353548d1b9ce21de3e22843a28c5df654f.jpg" }, { "user_id": 3212248, "full_name": "David Li", "title": "Founder", "founder_bio": "Building foundation models of the future @ Induction Labs", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/davidy-li/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fdfa320022a46b1bbdfc1ec63b6c4ca91cab0451.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107033, "title": "Induction Labs - foundation models that learn from observation", "tagline": "Unlocking scalable video pretraining with imagination models", "body": "**TL;DR: We're Induction Labs, and today we're announcing research we've been working on: imagination models, a foundation model architecture that unlocks scalable learning from internet video. Our first model, Photon-1, beats a production LLM on internal computer use benchmarks with 30x less training compute, at 3x lower serving cost. \\[[Main Article](https://inductionlabs.com/news/scaling-video-pretraining)\\]**\n\n![uploaded image](/media/?type=post&id=107033&key=user_uploads/2757445/94f0365c-7de2-4792-ad4e-a8362f27c233)\n\nInternet video contains millions of hours of people using computers, doing skilled work, and interacting with the world.\n\nImagination models can scalably learn from this video to get better at completing tasks and understanding the world.\n\nWe tested the architecture with Photon-1, a 106B-A5B MoE transformer pretrained on 18 years of computer screen recordings.\n\n**Some Results:**\n\n* **Outperforms a production LLM (Gemini 3.1 Flash-Lite) on our internal computer use benchmarks**, with 30x less pretraining compute and 3x cheaper inference.\n\n![uploaded image](/media/?type=post&id=107033&key=user_uploads/2757445/7e217b7a-bbf0-40b0-a160-94855bbf7e90)\n\n* **Simulates full desktop sessions from a single screenshot**: it can imagine VS Code, Gmail, and ChatGPT frame by frame\n* **Generalizes beyond computers**: finetuned, it learns checkers and simulates billiard physics better than LLM baselines.\n* **Picks up human behavior from video**, like reprompting ChatGPT until it gets what it wants.\n* [**Demos here**](https://inductionlabs.com/news/scaling-video-pretraining)\n\nWe see imagination models as a path to intelligence that learns by observing the world directly, without a human first translating it into text. \n\n**Our Ask**\n\nWe’re working to scale this method to more kinds of video and looking for exceptional people to work with us! If you’re an exceptional thinker, engineer, or researcher (or know anyone who would be a great fit) - shoot us a message at [team@inductionlabs.com](mailto:team@inductionlabs.com)", "url": "https://www.ycombinator.com/launches/RqL-induction-labs-foundation-models-that-learn-from-observation", "votes": 15, "created_at": "2026-07-23T16:02:32.265Z" } ] }, { "id": 30728, "name": "OS3", "slug": "os3", "former_names": [ "Observee" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/bd88161755dbb4d7513a2dc91e69b26930313524.png", "website": "https://os3robotics.com/", "all_locations": "San Francisco, CA, USA", "long_description": "OS3 is building the platform that makes intelligent robots accessible to every business.\r\n\r\nToday, robotics is constrained by two fundamental bottlenecks: hardware is too expensive to deploy at scale, and robots are difficult to train because high-quality robotics data is scarce and costly to collect. \r\n\r\nWe built OS3 to solve both.\r\n\r\nOur two-person team designed and a US-made mobile humanoid robot for a fraction of the cost of commercial alternatives. On the intelligence side, instead of relying on expensive robot-generated training data, we trained our own video-action model on 150,000 hours of real-world video spanning more than 2,000 tasks. Every deployed robot feeds new experience into our reinforcement learning pipeline, self improving robot fleet over time.\r\n\r\nThis creates a robot that can be deployed across industries to automate repetitive, physical work without requiring a new robot or training pipeline for every task. Our robots are already deployed in hotels, with customers now expanding the platform into biotech, hospitals, and data centers, including a wet lab at Stanford.\r\n\r\nOur mission is to do for robotics what Apple did for personal computing: take technology that was once expensive, specialized, and inaccessible and make it powerful and usable at massive scale. We’re building toward a world with billions of intelligent robots, and OS3 is the platform that makes that possible.", "one_liner": "Affordable, intelligent humanoid robots built to deploy at scale", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1753297896, "tags": [ "Hardware", "Robotics", "B2B", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/os3", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/os3.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/os3", "x_url": "https://x.com/os3robotics", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/bd88161755dbb4d7513a2dc91e69b26930313524.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 1427209, "full_name": "Rishabh Chanana", "title": "Founder", "founder_bio": "Built a computer at 10. First line of code at 12. Built & programmed a robot at 17.\n\nGrad School at UCSD, Robotics & AI. NLP research with Harvard and Georgia Tech. Trained LLMs at ServiceNow & built patented ML systems.\n\nNow soldering and training robots at OS3.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/risos8200/", "x_url": "https://x.com/risos8200", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5ef6e03c90f9e979491f4a8676962acf0df0db05.jpg" }, { "user_id": 1909563, "full_name": "Chris Hailey", "title": "Founder", "founder_bio": "Co-Founder at OS3. Ex-Coinbase software engineer on the asset addition team, generated over $19 billion of trading volume. Turned a YC hackathon winning project into a $10K MRR idea. Undergrad CS from USC with Presidential Merit scholarship. Serial marathon runner, most recently completed the 2025 Oakland Marathon in 3 hours 12 minutes.", "is_active": true, "linkedin_url": "https://linkedin.com/in/chrishailey8", "x_url": "https://x.com/realChrisHailey", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8ef8f471703ec1718bd79c1c35df1da5a33528f6.jpg" } ], "jobs": [ { "id": 100908, "title": "Mechanical Design Engineer - Robot Parts & Enclosures (Contract)", "url": "https://www.ycombinator.com/companies/os3/jobs/r2aKeWa-mechanical-design-engineer-robot-parts-enclosures-contract", "location": "San Francisco, CA, US / Remote (IN)", "type": "Contract", "role": "Engineering", "role_type": "Hardware", "salary_range": "₹360K - ₹800K INR", "equity_range": "", "experience": "Any (new grads ok)", "visa": "US citizenship/visa not required", "skills": [ "CAD Design" ] } ], "news": [ { "title": "Meet The YC Startups Betting On What Comes Next", "url": "https://www.forbes.com/sites/dariashunina/2026/09/03/meet-the-yc-startups-betting-on-what-comes-next/", "date": "Sep 03, 2026" } ], "launches": [ { "id": 110139, "title": "OS3: Intelligence and Machines for Physical Labor", "tagline": "Semi-humanoid hardware & Model based on 150k hours of Videos. Preorders Live Now", "body": "**TL;DR: OS3 is building semi-humanoid robots for physical labor. Our first robot, H.A.L.E 1.0, is built for everyday human tasks, and our first model is based on 2,000+ distinct real-world tasks and 150,000 hours of video data. We are opening up pre-orders to all with shipments starting this Fall.**\n\n**Physical AI Problem:**\n\nA lot of teams and businesses want to build and deploy in the robotics space. But they have to source parts or buy hardware from different vendors. Furthermore, the recent import ban makes it even harder for beginners and startups to use the latest hardware, unless they are willing to spend tens of thousands of dollars. Post hardware, constant data collection is needed to adapt open or closed-source models to the ever changing hardware choices. In other words, everyone keeps reinventing the same wheel of modern robotics development over and over again. \n\n**Our solution:**\n\n[**Video Link**](https://youtu.be/dDwaQbZYPLQ)\n\nWe built OS3: Both the hardware from ground up and the model to control the hardware. Our first robot, H.A.L.E. 1.0 (Human Aligned Life Enhancing), is designed to handle a wide range of everyday physical tasks, and we are presenting the prototype unit today.\n\nOur goal is towards pushing American robotics forward while ensuring the state of the art is available to all. \n\n**Hardware and Model:** \n\nIt features a bimanual embodiment with an omni-directional base, full height control to pick up objects directly from the ground while being able to reach top shelves, and a battery that can last a whole work day. Additionally, our design is primarily focused on safety and user-friendliness; a person can start and stop a task at just the touch of a screen on the back of the robot or a mobile app. \n\nThe intelligence layer is a two system architecture, with a high level planner layer and a lower level execution layer. The model has been trained on 150,000 hours of video data spanning over 2,000 real world tasks. Technical report coming soon, once evaluation runs finish. Hardware specs and engineering deep dive posted on our website.  \n\n**Preorders Live now** \n\nShipments begin Fall 2026.\n\nOur strong supply chain enables us to make our first low-cost robot starting at just under $10k for businesses to deploy (available in private beta now) and under $15k for developers to experiment with. We are already working with hospitality providers and biotech companies to deploy our robots. \n\n[**Order Link**](https://os3robotics.com/deploy/)**: 99 dollars deposit fully refundable**\n\nWe are also working on our own custom actuators right now, enabling even higher payload for our robots. Happy to discuss more if your business involves such use cases.\n\n**Talk to us:** [hi@os3robotics.com](mailto:hi@os3robotics.com) or book some time [here](https://cal.com/os3robotics/bf?overlayCalendar=true). ", "url": "https://www.ycombinator.com/launches/SeR-os3-intelligence-and-machines-for-physical-labor", "votes": 7, "created_at": "2026-08-11T19:00:00.126Z" } ] }, { "id": 30735, "name": "Evergrove", "slug": "evergrove", "former_names": [ "No official name yet", "Nony AI", "Evergrove Labs" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/bbb31241cfe39c763acf55ec7b3027f6e806380c.png", "website": "https://www.evergrovelabs.com/", "all_locations": "New York City, NY, USA", "long_description": "We've built voice agents for the largest work comp managed care orgs and specialist PT networks in the country, making >500K calls annually and accelerating the speed at which they coordinate care. On average, our customers see a 40% increase in workflow completion, 10x decrease in cycle times, and >60% reduction in cost.\r\n\r\nWe think voice agents are the starting point for solving slow and fragmented care coordination in this market. Companies that manage an injured worker's care rely on short-staffed contact centres and archaic methods of info exchange: phone, fax, and email. Every delay along the path of care leads to slower rehab of the injured worker, lower throughput for providers, and higher expense for the insurer/employer.", "one_liner": "Instant healthcare for all, starting with work comp", "team_size": 2, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1786637393, "tags": [ "Healthcare", "Insurance", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/evergrove", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/evergrove.json", "year_founded": 2026, "city": "New York City", "country": "US", "linkedin_url": "https://www.linkedin.com/company/evergrovelabs", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/bbb31241cfe39c763acf55ec7b3027f6e806380c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harj Taggar", "url": "https://www.ycombinator.com/people/harj-taggar" }, "founders": [ { "user_id": 1081614, "full_name": "Sid Unnithan", "title": "Founder", "founder_bio": "Co-founder and CEO. We're building interoperability in healthcare and commercial insurance. Voice agents are our bread & butter. Prev AI PM lead at Newfront (Acq for $1.3B), Azure ML & VS Code at MSFT, Waterloo Eng Alum (2018), and co-founder of appinkflow.com (bootstrapped).\n\nCo-founder and I built AI-native placement service within Newfront. Small team of expert humans and agent swarms negotiated policies between $300K and $1M USD, achieving 40% faster renewal times, and 15-20% cost reduction.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/siddhanthunnithan", "x_url": "https://x.com/SidUnnithan", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b219010304c97621f212e3f6017cac88384bced1.jpg" }, { "user_id": 1161116, "full_name": "Jacob Chia", "title": "Founder/CTO", "founder_bio": "Co-founder and CTO. We build agents for healthcare enterprises. Prev AI engineer at Newfront (Acq for $1.3B). Was on the platform team, wrote code for millions of authentication and authorization interactions at Newfront; Refactored file management for millions of mission critical insurance files.\n\nCofounder and I built fully autonomous insurance placements in 2025 - negotiating insurance coverage over email for policies between 300k to 1 million USD, achieving 15-20% cost reduction.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jacob-chia-78a1b2135", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9246fa7dc3b92b39196a646c9b93b9fe22a79e7a.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112258, "title": "Evergrove: Interoperability for work comp", "tagline": "Voice agents that accelerate post-claim care", "body": "We’re keeping this short and sweet. We build voice agents to overcome slow and fragmented care coordination in workers’ comp. Some key points:\n\n1. We’re on track to make >500K calls annually, helping customers get to >60% reduction in cost and 40%+ improvement in workflow completion.\n2. Our customers are some of the largest managed care orgs, specialty networks, and ancillary service providers in the work comp space.\n3. We’re ex-Newfront, where we pioneered agents that were cited in the $1.3B acquisition [press release](https://www.wtwco.com/en-us/news/2025/12/wtw-to-acquire-newfront-a-specialized-broker-combining-deep-expertise-and-cutting-edge-technology). We’ve spent our careers inside commercial insurance.\n\nOur mission is to have Evergrove agents powering comms between **every** entity on the post-claim care pathway. If you want to learn more about how we’re seeing early network effects, reach out at [sid@evergrovelabs.com](mailto:sid@evergrovelabs.com).\n\n_“If I had more time I would have written a shorter \\[launch post\\]”_", "url": "https://www.ycombinator.com/launches/TCc-evergrove-interoperability-for-work-comp", "votes": 30, "created_at": "2026-08-26T17:30:00.291Z" } ] }, { "id": 30904, "name": "Tensr", "slug": "tensr", "former_names": [ "Core", "Tensr", "Core" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/7f3743327b1fc85ded6caddbd532e0c600eb2cb0.png", "website": "https://tensr.com/", "all_locations": "Emeryville, CA, USA", "long_description": "Robotic factories that build robotics in the US. Currently operating a 12,000 sqft factory in the Bay Area and actively manufacturing customer orders for some of the largest robotics companies in the US and the International Space Station.\r\n\r\nWe’re a group of Berkeley graduate robotics researchers who previously won a full scale autonomous IndyCar competition at 160mph.", "one_liner": "Robotic factories that build robots", "team_size": 7, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1763065446, "tags": [ "Deep Learning", "Hardware", "Robotics", "Automation", "Industrial" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/tensr", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/tensr.json", "year_founded": 2025, "city": "Emeryville", "country": "US", "linkedin_url": "https://www.linkedin.com/company/tensrinc/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/7f3743327b1fc85ded6caddbd532e0c600eb2cb0.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Nicolas Dessaigne", "url": "https://www.ycombinator.com/people/nicolas-dessaigne" }, "founders": [ { "user_id": 2763027, "full_name": "Eric Berndt", "title": "Founder/CEO", "founder_bio": "Robotic factories that build robots in the US. Prev Robotics researcher @ UC Berkeley where I raced autonomous Indy500 cars at 160mph", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ekberndt/", "x_url": "https://x.com/ekberndt", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8136fef854af650d0b22ac083d6585e09121a8ed.jpg" }, { "user_id": 916548, "full_name": "Adith Sundram", "title": "Founder/CTO", "founder_bio": "Founder @ Tensr. Robotics, ML and Controls @ UC Berkeley", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/adithsundram", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/dd20d8f9a84f4b33da6805740257605ff2175494.jpg" }, { "user_id": 2763062, "full_name": "C.K. Wolfe", "title": "Founder", "founder_bio": "Founder, Tensr. CS PhD (AI, Robotics), UC Berkeley. Former Technical Program Manager, UC Berkeley robotics lab under Berkeley AI Research (BAIR). Secured $17M+ in DARPA and industry sponsorship across Stanford and UC Berkeley via AI Racing Tech (2024). Team Manager during top-ranked U.S. league season. Editor-in-Chief of UC Berkeley BAIR’s official AI research publication (BAIR Blog).", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ckwolfe/", "x_url": "https://x.com/ckwolfeofficial", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9656e4b1ee0b0c58c5a7df8df5bcb18614685e68.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112228, "title": "Tensr - Robotic Factories that Build Robots from Humanoids to Robots for the International Space Station", "tagline": "Robotic Factories that Build Robots", "body": "Robotic factories building robots. Today we have a 12,000 sqft factory in the Bay Area manufacturing everything from full humanoids to robots for the International Space Station.\n\nOur thesis: Revitalizing domestic manufacturing means driving costs down through automation. We are building an autonomous factory that is able to autonomously process a new order end to end from prototyping to large batch production. That means supply chain management, design for manufacturing, industrial machine retooling, robotic adaptation, quality inspection, packaging, and logistics all automated. Goal: Create a factory that can autonomously go from placed customer order to completed unit scale manufacturing run in 24 hours.\n\nWe are currently manufacturing production runs from top labs, full humanoids, and even have a partnership to manufacture robotics for the International Space Station. We've also built supply chain relationships in electronics, actuators, SMT, computing, pneumatics, and raw material suppliers.\n\nA snippet of some of our machines in the factory:\n\n![uploaded image](/media/?type=post&id=112228&key=user_uploads/2763027/777065e0-e83d-4501-8799-8c490e883851)\n\n![uploaded image](/media/?type=post&id=112228&key=user_uploads/2763027/30775e59-e171-4949-b959-3dc06ffa5322)\n\n## **Manufacturing at the Core**\n\nThis isn't just a robotic manipulation company. Its a manufacturing company.\n\nInternally we have pipelines autonomous building large databases of actuators, and raw tracking and trying to predict what items will need to order now to meet our next day targets.\n\n## **The Harness**\n\nThe entire factory is a data harness we can \"backprop\" learnings through from errors in the customer designs, failed supply chain prediction, slow retoolings, to inaccurate robotic manipulation. Our core belief is that robotic machine learning should not rely on human data to scale but instead learn to improve from its own mistakes in the real world in real time. This closes the manufacturing data gap that currently limits large-scale factory automation.\n\n![uploaded image](/media/?type=post&id=112228&key=user_uploads/2763027/2baf7c74-80e3-49bb-a6bf-712ce0fb30fc)\n\n![uploaded image](/media/?type=post&id=112228&key=user_uploads/2763027/4ed0acb3-e1bc-4210-b4d0-b3f3614bc8da)\n\n## **Any Robot Anywhere**\n\nOur thesis is simple. We should be able to make an robot for any environment whether that be a humanoid for the factory floor or a robot for Low Earth Orbit. We currently have a partnership with the ISS to manufacture a robots for the space station.\n\nNASA Houston, Johnson Space Center\n\n![uploaded image](/media/?type=post&id=112228&key=user_uploads/2763027/e4558cfa-cf98-447a-acaf-7e83650e73a7)\n\n## **Team**\n\nAt heart we're a team of Roboticists. Before Tensr we were the leads for a Berkeley lab's autonomous racing team where we raced autonomous full-scale IndyCars at over 160mph as the ML, Sim, and Robotic Controls leads.\n\n![uploaded image](/media/?type=post&id=112228&key=user_uploads/2763027/c3772faa-f515-4ee0-b32e-2e586b3ae8e3)\n\nWe previously worked in industrial automation, aerospace manufacturing, medical device manufacturing, Berkeley AI Research (BAIR), physics simulation, and secured $17M+ in DARPA / industry sponsorship for UC Berkeley / Stanford research.\n\nOur founding engineers range from Berkeley Robotics PhD, ex-NASA, to 10+ years of manufacturing experience.\n\n## A Better World\n\nOn a personal note, not only do I believe we have the opportunity to build a great business but also to make a positive impact on the world. Automate away environmentally harmful materials, eliminate waste 0.1% at a time, and drive steady-state COGS toward only the fundamental inputs (materials and electricity), unlocking strong margins while collapsing the price of everyday goods / cost of living.\n\nLets build a better world :)\n\nA full view of our vision here:[ tensr.com/whitepaper](https://www.tensr.com/whitepaper)", "url": "https://www.ycombinator.com/launches/TC8-tensr-robotic-factories-that-build-robots-from-humanoids-to-robots-for-the-international-space-station", "votes": 6, "created_at": "2026-08-25T21:00:00.441Z" }, { "id": 95506, "title": "Tensr: Building Autonomous Robotic Factories", "tagline": "Building fully autonomous robotic factories that make scaling hardware as effortless as scaling on AWS", "body": "We're building fully autonomous robotics factories at Tensr to create the AWS of manufacturing. Our factories will receive on-demand orders from customers and spin up robots to operate industrial machinery, package, and ship the finished product all in 24 hours.\n\n![uploaded image](/media/?type=post&id=95506&key=user_uploads/2763027/704a7eb3-1ffc-4fc3-b806-32f3fd256418)\n\n[**https://youtu.be/HLijIfoDfpA**](https://youtu.be/HLijIfoDfpA)\n\n**Problem**\n\nMoving fast in manufacturing today doesn't exist. Iterating designs for manufacturing takes months of back and forth getting samples from China and pre-ordering manufacturing capacity several months before the holiday rush, only then to wait several weeks for shipping praying there isn't a strike at the docks. At the same time we see the complete collapse of America industrial capacity as American workers and experience leave the field and contracts are sent overseas.\n\n**Solution**\n\nWith the robotics revolution, for the first time it is possible to fully reimagine the factory floor - moving from rigid lines to modular stations, even in high mix settings. However, we believe that automating manufacturing is more than just deploying a foundation model or a humanoid into an existing factory.\n\nFactories need to be designed from the ground up _around_ robots, not the other way around.\n\nOver the past few weeks we turned our living room into an autonomous factory floor, with stations for CNC, moving stock, and a focus on assembly. After demo day, we'll expand to a full manufacturing facility, and start supporting batch orders of all sizes from other companies throughout the country. Currently we also are in talks to provide our own assembly models to other distributors.\n\n![uploaded image](/media/?type=post&id=95506&key=user_uploads/2763027/706f70f4-47a7-4466-8281-5cf6d081da43)\n\n![uploaded image](/media/?type=post&id=95506&key=user_uploads/2763027/e6420b31-3e74-49d3-b751-c901c5379665)\n\n**Team**\n\nWe're a battle-hardened team of UC Berkeley graduate robotics researchers (from Berkeley AI Research (BAIR)) who met in a lab, previously worked on and managed multiple DARPA contracts and led an autonomous IndyCar team where we raced full scale autonomous Indycars at 160mph (250kmh) as the Machine Learning, Simulation/Team Manager, and Robotics Control leads respectively.\n\n\\- [Eric Berndt](https://x.com/ekberndt) - MS/BS EE/CS, UC Berkeley, CoFounder + CEO\n\n\\- [Adith Sundram](https://www.linkedin.com/in/adithsundram/) - MS/BS ME, UC Berkeley, CoFounder + CTO\n\n\\- [C.K. Wolfe](https://x.com/ckwolfeofficial) - Current PhD AI/Robotics, UC Berkeley, CoFounder\n\nIn addition, our team brings direct operational experience from Rockwell Automation (industrial automation), aerospace manufacturing, and FDA-regulated medical-device production\n\n![uploaded image](/media/?type=post&id=95506&key=user_uploads/2763027/f907544d-cb70-45d2-b043-708a6739e40c)\n\n**Our Ask**\n\nWe're offering early manufacturing / prototyping pilots for hardware companies of all sizes as the new year starts and we open our first facility. If you or anyone you know fits in this bucket, reach out to [**founders@tensr.com**](mailto:founders@tensr.com)!\n\n![uploaded image](/media/?type=post&id=95506&key=user_uploads/2763062/4e53c620-4423-4e57-b0cf-a25b750d0539)\n\n", "url": "https://www.ycombinator.com/launches/OqQ-tensr-building-autonomous-robotic-factories", "votes": 365, "created_at": "2025-11-14T01:26:45.320Z" } ] }, { "id": 30914, "name": "Sunflower", "slug": "sunflower", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ae86a0721b6effe108baf0067f19c1305be5c48b.png", "website": "https://sunflowerclinic.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Sunflower is building an AI-native telemedicine clinic for addiction. There are over 18,750 patients on a waitlist for Therapy or Meds, representing $153M worth of TAM just waiting to be served. \r\n\r\nWe started by launching Sunflower Sober, the first AI Sponsor for addiction which grew to 100 MAU in less than 6 months and currently driving about $1.5M in run-rate worth of subscription app revenue. \r\n\r\nSunflower Sober is currently pacing to add over 800,000 new users a year which means we are growing literally faster than Alcoholics Anonymous. \r\n\r\nWe have built one of the largest and most deeply loved brands within sobriety, and now we are building the clinical services underneath this brand to better fulfill on our mission & serve Medicaid who spends over $30B a year on SUD services. \r\n\r\nSunflower is founded by Koby who previously led Growth at Rupa from $5M to $75M/year. Our team includes clinical leaders (MDs / LPCs) from the Mayo Clinic, Choose Your Horizon, & Charlie Health. ", "one_liner": "We are building an AI-native telemedicine clinic for addiction. 🌻", "team_size": 15, "industry": "Healthcare", "subindustry": "Healthcare -> Healthcare Services", "launched_at": 1757558729, "tags": [], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare", "Healthcare Services" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/sunflower", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/sunflower.json", "year_founded": 2024, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/in/kobyconrad/", "x_url": "https://x.com/SunflowerSober", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ae86a0721b6effe108baf0067f19c1305be5c48b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Garry Tan", "url": "https://www.ycombinator.com/people/garry-tan" }, "founders": [ { "user_id": 132378, "full_name": "Koby Conrad", "title": "Founder/CEO", "founder_bio": "CEO & Founder @SunflowerSober 🌻 we're on a mission to create one billion days sober, one day at a time. YC S19 alumni, angel investor, growth marketer.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/kobyconrad/", "x_url": "https://x.com/kobyjconrad", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/34384404a013436d81dbaf00896dfb14f2ffe0fc.jpg" } ], "jobs": [ { "id": 116216, "title": "AI SEO Lead", "url": "https://www.ycombinator.com/companies/sunflower/jobs/EJZyKD7-ai-seo-lead", "location": "Buenos Aires, Buenos Aires, AR / Remote (US)", "type": "Full-time", "role": "Marketing", "role_type": null, "salary_range": "$1.5K - $12K / monthly", "equity_range": "0.01% - 0.10%", "experience": "1+ years", "visa": "US citizenship/visa not required", "skills": [ "Search" ] } ], "news": [], "launches": [ { "id": 112046, "title": "The Sunflower Clinic of Addiction Medicine (live in PA & MI) 🌻", "tagline": "We're building a 50 state telemedicine program for SUDs, reimbursed by medicaid.", "body": "**TLDR:** Sunflower Sober started by building the best AI Sponsor on the market & powering over 21,500,000 days of sobriety. Now we are launching the **Sunflower Clinic of Addiction Medicine** to bring Therapy & Medication to our users, reimbursed by medicaid. We have over $153M worth of patients on a waitlist.\n\n**App:** \n\n**Clinic:** \n\n![uploaded image](/media/?type=post&id=112046&key=user_uploads/132378/fc76c836-d166-4caf-b46c-fc2f85ddb653)\n\nHello friends!!\n\nDid you know that it is harder to build a **50 state SUD outpatient program that takes medicaid**, than it is to put rockets into space?\n\nThe licensing requirements once you start treating substances as a primary diagnosis are quite frankly insane, state-level medicaid approval, county level MCO contracts, physical facilities in every state.\n\nBut we started Sunflower to solve **hard problems**.\n\nMost people look at us and think “hehe, a mobile app, a sobriety counter”. When we raised our pre-seed with no product and no revenue I told our early investors that this is my version of going to mars. Sunflower isn’t here to just build a sobriety counter. **We are trying to literally and objectively end addiction.**\n\nWe’re on a mission to create one TRILLION days of sobriety, one day at a time.\n\nSo far we have powered over 21,500,000 days.\n\nWe built the first AI Sponsor, named Sam, who has now had over **7,000,000 messages** with our users. More messages than any “entity” has ever had within a single lifetime about sobriety.\n\nSam is a critical piece to the sobriety journey. Almost 80% of surveyed users at d60 report Sunflower preventing at least one relapse, and **93.6% report a reduction in substance usage**.\n\nBut we’ve always known that an AI Sponsor is only a PART of what people need to get sober. They still need **therapy**. They still need **medication**.\n\nThe problem?\n\nWhen you are struggling with substance abuse (alcohol or otherwise), you often don’t have $10 in your bank account to pay for gas.\n\nWe have too many users that downloaded Sunflower because **they can’t afford paying $125/hr for therapy**. That can’t afford dropping $350 a quarter for Naltrexone.\n\nOutpatient programs for SUDs are one of the most regulated spaces within behavioral health, and because of this, access to care that _will actually take our users insurance_ (medicaid) is extremely limited.\n\nThis is why we are working to build the **Sunflower Clinic of Addiction Medicine**.\n\nA 50 state telemedicine program for SUDs that _actually_ takes medicaid. Now live in MI & PA with SUD facility licensing, state medicaid approval, and in the final steps of county level MCO contracting.\n\nWe are combining our best in class AI Sponsor with Mayo Clinic MDs and therapy leadership that built group programs at Charlie Health from 8 - 300+ therapists … our goal is to build the **best outpatient program on the face of this planet** & get it covered by medicaid nationwide.\n\nSo far we have over **37,500+ patients on a waitlist** who have requested Therapy or Medication from Sunflower, representing a TAM of $153M waiting to be served.\n\nWe are going to solve one of the most important problems in human existence. A problem that globally we spend over $100B treating every year, and is the greatest shackle to the acceleration of humanity.\n\n**Thanks for following along with our story.** Here’s some data. 🌻\n\n![uploaded image](/media/?type=post&id=112046&key=user_uploads/132378/bc43bb69-3c76-4b16-8e62-9cfb56adef66)\n\n![uploaded image](/media/?type=post&id=112046&key=user_uploads/132378/6360a175-3730-4f34-8f8f-877c0ecab8e8)\n\n![uploaded image](/media/?type=post&id=112046&key=user_uploads/132378/811dae90-30c2-433d-93fb-876fb9f2beb3)\n\n![uploaded image](/media/?type=post&id=112046&key=user_uploads/132378/8202d2c7-82ba-45c5-bf9b-c9c2373454b5)\n\n![uploaded image](/media/?type=post&id=112046&key=user_uploads/132378/d23f1ef9-74b8-4e82-8020-6b8938e6acc5)\n\n![uploaded image](/media/?type=post&id=112046&key=user_uploads/132378/96db8f55-0afb-4cf5-bb78-715ba7e29fa6)\n\n![uploaded image](/media/?type=post&id=112046&key=user_uploads/132378/f3fec3da-c919-4264-8946-5aed579c3493)\n\n", "url": "https://www.ycombinator.com/launches/T9C-the-sunflower-clinic-of-addiction-medicine-live-in-pa-mi", "votes": 526, "created_at": "2026-08-24T16:12:24.013Z" }, { "id": 94154, "title": "Sunflower: An AI Agent tasked with helping you stay sober. 🌻", "tagline": "Quit drinking alcohol, nicotine, cannabis, opioids, heroin, gambling, porn, or anything else", "body": "![uploaded image](/media/?type=post&id=94154&key=user_uploads/132378/03768fda-8234-42c4-a4ef-07df5eccdb39)\n\nHey friends!\n\nSunflower builds an app that **uses AI to convince people to stop drinking** (or doing drugs) & **to stop killing ourselves**. 🐝\n\nMost people have a personal story or connection with addiction.\n\nFor me, I used to smoke too much weed (thank god it wasn’t anything harder). My grandpa passed away from alcoholism & my grandma passed away from lung cancer.\n\n“Addiction” is a supply constrained market, which means that there isn’t enough supply (good doctors) to properly service the 1 billion+ humans who struggle with addiction.\n\nBut because of an AI, **for the first time EVER**, we have an infinitely scalable solution to effectively treat addiction.\n\n**Sunflower grew from 200 → 100,000 Monthly Active Users in less than 6 months**, and now we’re in almost every country on earth in English, Spanish, Japanese, Korean, Portuguese.\n\nSorry there’s no “video” for this Launch YC. Best way to experience our product is to simply [go download it](https://sunflowersober.com/).\n\n**Here’s the thing i’m most proud of**, which is what our users actually say about this thing that we built:\n\n![uploaded image](/media/?type=post&id=94154&key=user_uploads/132378/02950d8c-fc04-4f17-93be-075ed219c1ba)\n\n![uploaded image](/media/?type=post&id=94154&key=user_uploads/132378/97b08723-3fa7-493f-b1cc-9733d3270220)\n\n![uploaded image](/media/?type=post&id=94154&key=user_uploads/132378/fc3432ff-3c0b-48af-b989-0cba27824309)\n\n![uploaded image](/media/?type=post&id=94154&key=user_uploads/132378/8bfa5024-b885-4b80-a532-fe88966d4b01)\n\n![uploaded image](/media/?type=post&id=94154&key=user_uploads/132378/3a18de92-bd65-4294-9ea5-73d7a140ba46)\n\n![uploaded image](/media/?type=post&id=94154&key=user_uploads/132378/37117421-e206-4e2e-b07f-20aac23d3455)\n\n![uploaded image](/media/?type=post&id=94154&key=user_uploads/132378/6b7ea32c-b46d-4fd7-a046-2577d9102a71)\n\n", "url": "https://www.ycombinator.com/launches/OUc-sunflower-an-ai-agent-tasked-with-helping-you-stay-sober", "votes": 456, "created_at": "2025-09-26T20:58:53.448Z" } ] }, { "id": 30923, "name": "Frontier Computing", "slug": "frontier-computing", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/a10237e5548f0937c67885cb916abadcaa74cc08.png", "website": "https://frontier.site", "all_locations": "London, England, United Kingdom", "long_description": "Frontier is first to grow biological brains at scale, serving biological neural nets with substrate-level colocation of memory and compute, enabling highly cost effective training of large parameter count models.\r\n\r\nWhile academia has traditionally been limited to ~1M neurons in vitro, Frontier has built a pathway to scale to 100M+ neurons by circumventing traditional scaling challenges in biological culture, demonstrating this strategy's functionality by training a biological brain using this scaling primitive to play Frogger.\r\n\r\nFrontier is now growing exascale compute clusters to produce substrates capable of long-horizon language reasoning while operating at much higher sample efficiencies than traditional architectures, and collocating memory with compute to enable significantly larger parameter counts than would be tractable for training within traditional compute clusters.", "one_liner": "Frontier grows scalable biological brains as an ML training substrate", "team_size": 3, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1788161513, "tags": [ "Hard Tech", "Hardware", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United Kingdom", "Europe" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/frontier-computing", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/frontier-computing.json", "year_founded": 2025, "city": "London", "country": "GB", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/a10237e5548f0937c67885cb916abadcaa74cc08.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 2627752, "full_name": "Michael Domarkas", "title": "Founder/CEO", "founder_bio": "19 y/o growing exascale brains for continual learning\n\nprev growing brains in my bedroom", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/michael-domarkas", "x_url": "https://x.com/MichaelDmrks", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/dbc0487122f2a39f809799e732a7b1e3cb67b540.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112791, "title": "Frontier Computing – Growing biological brains at scale", "tagline": "We colocate memory with compute in biological tissue, so you can train larger models at lower cost", "body": "TL;DR – We're deploying exascale biocomputing clusters, growing biological systems with memory baked into compute at the substrate level, enabling training of models at parameter counts that would otherwise be inaccessible under traditional GPU architecture-constrained approaches.\n\n![uploaded image](/media/?type=post&id=112791&key=user_uploads/2627752/445c283b-feb2-4b6b-9d2c-ab2cf60edc7d)\n\nHey all, we're Frontier Computing, and we’re excited to announce that we are growing the first large-scale neuronal tissues for ML training and inference!\n\nFrontier is launching its 500M biological neurone cluster by January 2027, which will be 2500x larger than the next-largest publicly available system in the biocomputing space that companies are currently restricted to building on. \\\n\\\nWe'll continue to launch increasingly larger iterations of our substrate to circumvent the challenges of GPU parallelisation seen in roofline analysis across computation/computation/memory, colocating memory with compute in biological neurones to circumvent the infra problem of GPU networking for model training and inference, enabling inference time training at scale on biocompute, at much lower cost than traditional GPU approaches.\n\n# **❌** The Problem\n\nBiological brain tissue is already being bought and sold, including being trained for a range of tasks from gaming environments to next-token prediction, but the scaling limits of the underlying wetware have acted a fundamental barrier to useful economic output.\n\nAcademia has previously been unable to exceed a \\~1M neurone scale due to \"the vascularisation problem\", where neurones cannot grow past a geometric limit imposed by insufficient oxygen access in cells that are increasingly far away from active supply.\n\n# **✨** Our solution\n\nFrontier has developed a novel neuronal tissue culture approach to engineer large-scale brain systems that exceed traditional scaling limits of neuronal tissue culture.\n\nThis enables us to scale past traditional wetware substrate sizes, supporting much larger biological neural nets with correspondingly greater parameter counts, increasing the complexity of task environments these systems can represent, and using their larger geometric size to provide physically larger surfaces for I/O and informationally rich encode and decode.\n\nWe provide the substrate, interfacing and cloud infrastructure to enable third parties to train on our systems, taking advantage of the inference-time training capacities of biological neurones, the sample efficiency, the capacity to exist in an abstraction layer above backprop, and the opportunity to perform RL on models in excess of sizes that would be traditionally accessible on GPU hardware due to memory bandwidth issues that we circumvent through colocating memory and compute within neuronal tissues.\n\n## **🏆** What we've achieved so far\n\n\\- We have proven out the first scalable culture system for biocompute, and demonstrated its computational capacities by training our system to play the arcade game Frogger\n\n\\- Demonstrated tail-end capabilities in sample efficiency, with 1 hour of real-time learning producing 92% peak cross rate (over a 25-game sample) in our Frogger environment\n\n\\- Raised a 10M pre-seed round to produce early scale up our team and wetlab work, led by General Catalyst with participation from LocalGlobe, Amino Collective, Kaya, Long Journey Ventures, Off Piste Capital, SVA and more angels and firms that we are very excited to have on board\n\n\\- Begun build out of a 100M and a 500M biological neurone system, live EOY 2026\n\n## **🔜** What's next\n\n\\- Demonstrating SOTA task performance of biohybrid models using our larger-scale substrates\n\n\\- Launching our API library to enable our customers to implement their own models and RL environments via our platform\n\n\\- Iterating towards post-neocortex-scale build outs to further enhance computational capacities of biocompute\n\n## **👥** Our Team\n\n[Michael](https://www.linkedin.com/in/michael-domarkas/), CEO, is a Cambridge Natural Scientist who first learnt programming at 8 through C# Unity game development, competed in maths/chess/bio competitions nationally throughout high school, and has been culturing neuronal tissues for the past two years, starting from $20K USD in grant funding at 17 from Emergent Ventures. He has built the first scalable neuronal culture platform for biocomputation and validated performance via reinforcement learning on this substrate.\n\n## **🙏** Asks\n\nJoin the team! We're hiring across the stack for neurobiologists, RL researchers, electrical engineers, and biochemists, accepting both direct applications and referrals. Feel free to get in touch at the link below:\n\n\n\nIf you're interested in working at the _frontier_ of biocomputing and scaling the fundamental wetware primitives humanity has access to, Frontier Computing is building out the leading platform in the space. We’re looking forward to collaborating with you.", "url": "https://www.ycombinator.com/launches/TLD-frontier-computing-growing-biological-brains-at-scale", "votes": 21, "created_at": "2026-08-31T17:00:00.185Z" } ] }, { "id": 31047, "name": "Waddle Labs", "slug": "waddle-labs", "former_names": [ "Waddle" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/f525546cff510b464f7877004fcc10c8f0b97843.png", "website": "https://www.waddlelabs.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "", "one_liner": "LLMs that control robots", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1784890498, "tags": [ "Developer Tools", "Generative AI", "Robotics", "Automation" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/waddle-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/waddle-labs.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/waddlelabs/", "x_url": "https://x.com/theWaddleLabs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/f525546cff510b464f7877004fcc10c8f0b97843.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 2132817, "full_name": "Hanming Ye", "title": "Founder", "founder_bio": "Interested in how natural intelligence arises. I also like machines such as 3D printers and robots. \n\nhttps://dozenducc.github.io/", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/hanming-ye-0467b01aa", "x_url": "https://x.com/DozenDucc", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5b813f423ca4b4732c54817689193d5602190bf5.jpg" }, { "user_id": 1853600, "full_name": "Yiding Song", "title": "Founder", "founder_bio": "Interested in systems that learn theories about the world with human-like efficiency.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/yiding-song-vincent/", "x_url": "https://x.com/yiding_song", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b7078e0f6f101f74db1f4584bb6cd102051bb9f0.jpg" } ], "jobs": [ { "id": 109947, "title": "Founding Research Engineer", "url": "https://www.ycombinator.com/companies/waddle-labs/jobs/RaRUAQK-founding-research-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Machine learning", "salary_range": "$175K - $350K", "equity_range": "1.00% - 2.00%", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [ "ML", "Python", "Torch/PyTorch", "Distributed Systems", "Fine-tuning", "Evals", "AI Agents", "Diffusion Models" ] } ], "news": [], "launches": [ { "id": 107821, "title": "Waddle Labs - Agents that control robots", "tagline": "Claude Code for robotics - prompt your robot to do physical tasks", "body": "Waddle is the Claude Code for robotics. Connect our API to your robot and type a prompt, then our agents write code to achieve the task in \\~20 minutes.\n\n![uploaded image](/media/?type=post&id=107821&key=user_uploads/2132817/afc74f28-33c6-450c-80c3-7f2d263f213d)\n\n### **Our Hypothesis**\n\nRobot intelligence today depends either on large end-to-end VLAs or classical approaches.\n\n\\- VLAs are difficult to customise and steer, require huge amounts of data to train, and don’t yet generalise across embodiments and environment. They face limited deployment.\n\n\\- Meanwhile, classical approaches are widely deployed for their reliability and explicit control, but are often brittle and costly to redevelop in new setups.\n\n_Does this tradeoff have to be inevitable?_\n\n**Our hypothesis:** By using agents to control robots, we combine the generalisation of VLAs with the speed and reliability of classical approaches. \n\n### **Our Solution**\n\nOur agents decomposes a goal into subtasks, and completes each subtask by viewing camera feeds, writing control code, and calling models like VLAs. The agent outputs a program that you can run and iterate on by talking to our agents. \n\n**Where we've been using Waddle internally:**\n\n\\- **Create a working policy in 20 minutes** (\"Write a program to package the box\")\n\n\\- **Generate data for model training** (\"Pick and place lego bricks at random positions 1000 times\")\n\n\\- **Facilitate robotics auto-research** (\"I am tuning a policy overnight. Reset the scene after each trial”)\\\n\\\nBecause our agents are LLMs, they inherit properties of the underlying models. Agents and the code they write are inherently generalist, so they work with any robot and any environment without additional data. Moreover, agents excel at long-horizon planning and can decompose a physical goal into subtasks, verify each stage, and re-plan on failure.\n\n**Interested in trying our agents via our preview API?** Sign up at \\[our website\\]() or email [founders@waddlelabs.ai](mailto:founders@waddlelabs.ai)\n\n### **Our Research**\n\nWaddle's agents enable new forms of scaling that we are very excited about:\n\n\\- **Scaling compositional generalization:** Our agent first wrote the “fold_grasp” skill while flipping a package; another agent later used it to fold a t-shirt. Over time, agents accumulate a library of skills, enabling future agents to compose them and achieve harder tasks.\n\n\\- **Scaling multi-robot cooperation:** Imagine an entire fleet of robots collaborating to achieve a task. For VLAs, this is a nightmare. For Waddle, simply hook up more agents. A master agent spawns subagents, coordinating multiple robots working concurrently.\n\n\\- **Scaling from scale:** We evaluated Opus 4.8, Fable 5, and GPT 5.6 Sol on a suite of manipulation tasks: larger models with larger thinking budgets consistently produce better policies. As foundation models improve, so do our robots.\n\n### **The Team**\n\nWe are @Yiding Song (Vincent) and @Hanming Ye. We met at MIT’s Research Science Institute, where Vincent was working on multimodal astrophysics models and Hanming was researching topology in the math department. Since then, we’ve become roommates at Harvard where we are either doing ML research or breaking robots - usually both.\n\nThere will be millions of robots in the future. Our mission is to enable everyone to take part in building them.", "url": "https://www.ycombinator.com/launches/S33-waddle-labs-agents-that-control-robots", "votes": 11, "created_at": "2026-07-27T17:45:25.918Z" } ] }, { "id": 31074, "name": "OneTriangle", "slug": "onetriangle", "former_names": [ "EvalAI", "TrustAI Labs", "TrustAI" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/75197a50573558e4b8c16932b7db7c7c11f235a4.png", "website": "https://onetriangle.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Inference has surpassed training as AI’s largest compute cost, accelerated by increasing demand for AI and agents. By using proprietary KV cache transfer tech, OneTriangle has cut inference costs by 20% and time by 40% compared to present standards.", "one_liner": "The cheapest, fastest inference cloud", "team_size": 4, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1783378352, "tags": [ "Open Source", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": [ { "question": "How did your company get started? (i.e., How did the founders meet? How did you come up with the idea? How did you decide to be a founder?)", "answer": "Hi, we’re Hannah and Medha! Three years ago, we met at MIT, and now are co-founding OneTriangle. Hannah loves scaling companies and Medha loves making ideas become real products. Together, we built OneTriangle to push the limits on latency, cost, and quality. " }, { "question": "What is your long-term vision? If you truly succeed, what will be different about the world?", "answer": "We believe the future of AI will be defined by intelligence and trust. AI agents will reliably carry out human intent while recognizing and refusing actions that are unsafe, unlawful, or harmful. With the rate of advancing ML systems, we will bridge this ever-increasing gap. We provide a compliance standard for all agents to follow, ensuring that humans and agents are perfectly aligned in their intentions." } ], "url": "https://www.ycombinator.com/companies/onetriangle", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/onetriangle.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/trustaiinference/", "x_url": "https://x.com/OneTriangleAI", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/75197a50573558e4b8c16932b7db7c7c11f235a4.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/f775cf723e5a0f2258a41df7227d4243e4b2b3b5.png", "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/d28c1c890fcf41f1e457a449b678b33930d05cb5.png" ], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 2573695, "full_name": "Hannah Chung", "title": "Founder/CEO", "founder_bio": "MIT CS & Econ → @Virtu Financial → Fastest, Cheapest Lightweight Inference @OneTriangle\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/hannahchung1/", "x_url": "https://x.com/hannah_chuu", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a97b9b67d35f8f68aeebc4b930d52ab56e6b4e49.jpg" }, { "user_id": 2043461, "full_name": "Medha Venkatapathy", "title": "Founder/CTO", "founder_bio": "Hi! I studied physics + cs at mit, and we're building the cheapest/fastest inference with KV cache transferability. In my free time, I like to bike and get lost.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/medhav/", "x_url": "https://x.com/medha_rv", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/543ebe65ad3407ffb95c2dd2e5183140b298be3f.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111487, "title": "OneTriangle - The fastest, cheapest inference, powered by KV cache transfer", "tagline": "Lightweight novel inference ", "body": "**TL;DR:** OneTriangle is building the cheapest, fastest inference. We cut inference prefill costs by 20% and time-to-first-token by 40%, letting you get large-model quality at small-model cost.\n\nVideo: \n\n**Team**\n\nWe’re Hannah (CEO) and Medha (CTO), MIT CSAIL grads building with a team of 4 MIT engineers spanning ex-DeepMind and Jane Street, Physics and Astronomy Olympiad medalists, and NeurIPS/ICML authors. \n\n**The Problem**\n\nInference is expensive. \n\n• You’re forced into a bad tradeoff. You need the decoding power of large models, but you have the budget for small ones.\n\n• Prefill is the hidden tax. For long-context and agentic workloads, prefill dominates cost and latency, and everyone just eats it.\n\n**Our Solution**\n\nOneTriangle transfers the KV cache between models, something that has never been done in production. We prefill on a small model, strip the positional encoding, map the cache into the large model’s space, and restore relative importances. The result:\n\n• 20% lower prefill costs\n\n• 40% faster TTFT\n\n• Large-model quality at small-model cost\n\nWe beat out NVIDIA’s KV cache transfer paper by getting a lower KL divergence score in 2 weeks.\n\n![uploaded image](/media/?type=post&id=111487&key=user_uploads/2573695/e64727d6-a7b1-404a-bfbb-22ad2b941142)\n\n\\\nNo API changes, no quality cliff, and we’re upstreaming it into vLLM so the fastest path to cheap inference is the infrastructure you already run.\n\n**Why Now**\n\nInference has surpassed training as AI’s largest compute cost, and it’s accelerating with demand for agents that generate 10x the tokens a chat session does. Inference already eats up to 90% of total compute cost, over $600 billion is going into AI data centers this year alone. Every team pushing the frontier is bottlenecked by the same thing: serving costs.\n\n\\\n**The Future**\n\nOur bet is that model-mixing becomes the default serving pattern: small models handle prefill and routine tokens, large models handle the hard ones, and the cache moves freely between them. Whoever owns the transfer layer owns the economics of inference. We are the future of lightweight, fast inference.", "url": "https://www.ycombinator.com/launches/T0B-onetriangle-the-fastest-cheapest-inference-powered-by-kv-cache-transfer", "votes": 10, "created_at": "2026-08-21T02:30:02.532Z" } ] }, { "id": 31086, "name": "Prescience, Inc.", "slug": "prescience-inc", "former_names": [ "HouseAI", "Prescience" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/9318ef0555c0ad2e5ce6049908b655658d073f1f.png", "website": "https://www.getprescience.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Prescience is the preeminent health benefits platform for the world's most ambitious companies. We save employers 20-50% on health insurance while offering best-in-industry health benefits: www.getprescience.com\r\n\r\nAbout us:\r\nPrescience is building a new healthcare system for America. We train AI systems to: create, operate, insure, and deliver healthcare for employer health plans. Prescience Time Machine – an AlphaGo-like system – steers patients towards “Move 37 health trajectories” that deliver dramatic cost savings for employers. Rishab Jain and Aditya Jain, M.D. are serial entrepreneurs and brothers from Harvard & Harvard Medical School with backgrounds in machine learning and medicine. We believe abundant intelligence will bring an opportunity to fundamentally rebuild the $4.5T healthcare industry around proactive (prescient) health and wellness.", "one_liner": "Superintelligent health insurance that keeps employees healthy", "team_size": 5, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1780947683, "tags": [ "B2B", "Health Insurance", "Insurance" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/prescience-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/prescience-inc.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/prescienceinc/", "x_url": "https://x.com/getprescience", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/prescience-2bc9", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/9318ef0555c0ad2e5ce6049908b655658d073f1f.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 1330533, "full_name": "Rishab Jain", "title": "Founder/CEO", "founder_bio": "Building a prescient healthcare system. Prev. CS/Neuro @ Harvard College", "is_active": true, "linkedin_url": "https://linkedin.com/in/rishab-jain-k", "x_url": "https://twitter.com/rishabjaink", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/57cc4755f0ec35b55ceb82a08fc87b934da21b5a.jpg" }, { "user_id": 3202469, "full_name": "Aditya Jain", "title": "Founder/President & COO", "founder_bio": "Building a prescient healthcare system. MD, Harvard Medical School.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/adityajain2000/", "x_url": "https://x.com/adityajain_42", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/92d44ac319ea99329f44e1c1f8d5098cd8810e81.jpg" } ], "jobs": [ { "id": 107013, "title": "ML Research Engineer", "url": "https://www.ycombinator.com/companies/prescience-inc/jobs/YcWt3KT-ml-research-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$250K - $350K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [ "TypeScript", "Machine Learning", "Deep Learning", "Next.js" ] }, { "id": 107167, "title": "Founding Engineer", "url": "https://www.ycombinator.com/companies/prescience-inc/jobs/O7WXOyJ-founding-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$180K - $250K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 107168, "title": "Founding Operations / Head of Operations", "url": "https://www.ycombinator.com/companies/prescience-inc/jobs/zS3Q0ob-founding-operations-head-of-operations", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "$180K - $240K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 114034, "title": "Founding GTM", "url": "https://www.ycombinator.com/companies/prescience-inc/jobs/9RA2c6M-founding-gtm", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "$110K - $235K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 106632, "title": "Prescience: Superintelligence for the American health system", "tagline": "A new health insurance that saves employers money and keeps employees healthy.", "body": "Hey YC, we're [Rishab Jain](https://www.linkedin.com/in/rishab-jain-k/) and [Aditya Jain, MD](https://www.linkedin.com/in/adityajain2000/). \n\nToday, we’re announcing Prescience 🔮: health insurance for the age of abundant intelligence. Prescience is taking on United, Blue Cross, and every other incumbent to build the first health insurance company that doesn’t suck.\n\n**TL;DR -** Prescience is a health insurance solution built for a world with AGI. We help cut premiums by 20%, offer $0 deductibles, include GLP-1s and other modern medicines, and have no hidden fees: we’re only paid when we save you money. We’re already helping some of America’s most ambitious companies save money, provide better benefits, and keep employees healthy so they can focus on their business. [www.getprescience.com](http://www.getprescience.com)\n\n[youtube.com/watch?v=PBqyj9zP4BY&feature=youtu.be](http://youtube.com/watch?v=PBqyj9zP4BY&feature=youtu.be)\n\n**The Problem**\n\nHealthcare is a $6 trillion industry in America. For most companies it's the second-highest line item on their P&L, with costs rising about 15% every single year.\n\nEmployers keep paying more and getting less. Deductibles climb and the best benefits get cut. Even employees with \"good\" insurance still end up with denied claims, surprise bills, and debt.\n\nNothing changes because the entire industry is built to take a cut of your premium. When care gets more expensive, all of the middlemen make more money. They couldn’t care less if your people get more healthy.\n\n**Our Solution**\n\nPrescience is an AI-native health insurance that flips the incentive. Get a quote without speaking to middlemen, and get better health benefits at lower cost. That adds up to roughly **20% lower premiums**, and **better-than-platinum health benefits** like $0 deductibles, no middlemen, no fees. For YC companies, we have a Deal running with Free Wearables and GLP-1s covered.\n\nHere's how we deliver. **Time Machine** is our reinforcement learning system trained on medical histories. The same way AlphaGo found \"Move 37,\" a move no human would have played that turned out to be brilliant, Time Machine searches millions of possible health trajectories for each member and finds the non-obvious path that keeps them healthiest. Then it steers care toward that path: the right screening at the right moment, the right behavior, before a small issue ever becomes a hospital bill. We’re running a separate waitlist for YC companies who want to try Time Machine for personal use: \n\n![uploaded image](/media/?type=post&id=106632&key=user_uploads/3202469/9ec81a8e-0484-4a24-873c-9aafa87d8124)\n\nTime Machine is paired with **Crystal**: our care navigation assistant in your pocket. As an MD, I’ve seen how patients struggle to manage the labyrinth of healthcare. Hours of phone calls and wait times. Pouring through confusion insurance documents. No more - Crystal can book visits for you, answer questions about your health, and pay your bills.\n\n![uploaded image](/media/?type=post&id=106632&key=user_uploads/3202469/6d96fb46-38d8-4b14-934d-c070c77b13b4)\n\nInstead of today’s health insurance experience that forces employers to talk to brokers for weeks just to get covered, employers on Prescience get a quote in days and a plan that automatically tracks savings and does filings. We don’t just sit around and collect money from premiums, but actually try to make your employees healthy.\n\n![uploaded image](/media/?type=post&id=106632&key=user_uploads/3202469/869b2f99-6114-43c4-b8a3-4630545de430)\n\nWe are a real, full-stack solution. We’ve built out and partnered to deliver quotes, banking, HSA, MSO-PC, legal, compliance, network, risk, and any other solutions, We have a national team of top doctors who we negotiate with directly - no middlemen - and who we are able to offer fast payment, clear patient histories, and no admin paperwork. They in turn offer us heavy discounts we pass on to employers.\n\n**The Team**\n\nRishab and Aditya are brothers and serial entrepreneurs. Rishab (CEO) studied CS & Neuro at Harvard, is an RSI alumni, won the International Science and Engineering Fair, and previously founded and exited an edtech company. Aditya (President/COO) earned his MD from Harvard Medical School, has been doing ML research since 2014, and is an Op-Med Fellow at Doximity.\n\nPrescience aims to rebuild the entire $6T American healthcare industry into one that is proactive and prescient, not reactive. Our vision is to be both payer and provider, as our proprietary AI and data becomes ready to take on all of medicine.\n\nReach us anytime at [founders@getprescience.com](mailto:founders@getprescience.com).\n\n**Our Ask**\n\n1. **If you’re a founder, get your quote:** at [getprescience.com](http://getprescience.com) and **use our YC deal on Bookface**! Free GLP-1 weight loss program, genetic testing, and white-glove support.\n2. **Intro us to enterprise:** Founders, Chief People Officers, Benefits Managers, Chief Medical Officers, CFOs, HRIS Platforms, or Brokers, especially at companies reviewing 2027 benefits!", "url": "https://www.ycombinator.com/launches/Rjs-prescience-superintelligence-for-the-american-health-system", "votes": 61, "created_at": "2026-07-21T14:07:46.343Z" } ] }, { "id": 31161, "name": "Rapidfolio", "slug": "rapidfolio", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/03c9e049e23bc7dcc8df7c9c46fd48f3e74846af.png", "website": "https://rapidfolio.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Rapidfolio helps banks and fintechs fully automate back-office workflows like fraud investigations, KYB/KYC and loan underwriting, with accuracy and auditability.\r\n\r\nBuilt by fintech founders and operators, Rapidfolio is on a mission to ensure banks and fintechs scale their customer base without adding another back-office head count.", "one_liner": "AI-native back-office for banks and fintechs", "team_size": null, "industry": "B2B", "subindustry": "B2B", "launched_at": 1784234137, "tags": [ "Fintech", "SaaS", "Enterprise", "Automation", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/rapidfolio", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/rapidfolio.json", "year_founded": null, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/rapid-folio/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/03c9e049e23bc7dcc8df7c9c46fd48f3e74846af.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/b5d2e5015ae907316abef45c424725ff7c1bd646.png" ], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 334623, "full_name": "Naveen Qureshi", "title": "Founder/CEO", "founder_bio": "Co-founder and CEO Rapidfolio, AI for fintechs and banks.\nPrev: co-founder & CEO of Sable (S19), a consumer fintech that I built, scaled to few hundred thousand accounts and sold.\nOn a mission to help banks and fintechs grow their customer base without adding another back-office headcount. Ex Capital One, Amazon, Target. Columbia Business School.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/naveen-qureshi", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/35caac9e4b831d4e37de25f7c777c368e27cc364.jpg" }, { "user_id": 2636714, "full_name": "Frederik Vanhevel", "title": "Founder/CTO", "founder_bio": "Co-founder & CTO at Rapidfolio (S26), where we use AI to automate the most complex financial processes with self-evolving, always-compliant, fully-auditable workflows no one else is building. Previously built two banks from scratch: founding engineer at Sable (S19, shipped in a record 3 months, pre-AI) and Seen, where I rebuilt the entire infra for the Snap Finance acquisition, built the credit decisioning engine and scaled to serve hundreds of thousands of customers.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/frederik-vanhevel/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/0da9e1bb6735ce1052d2b98eff490d3b95bfe76a.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 106986, "title": "Rapidfolio - AI Platform for Compliance, Operations, Risk & Fraud for Fintechs and Banks 🏦", "tagline": "AI for banks and fintechs", "body": "Hi everyone 👋\\\nWe're [Naveen](https://www.linkedin.com/in/naveen-qureshi) & [Frederik](https://www.linkedin.com/in/frederik-vanhevel/), fintech founders and operators building [Rapidfolio](https://rapidfolio.com/).\\\n\\\n**TL;DR:** Rapidfolio automates the most complex manual compliance, operations, and fraud processes that analysts perform at fintechs and banks today. We do this with 100% accuracy, speed, and auditability. We are previous founders and operators in fintech, and created Rapidfolio because we wished it existed when we were running Sable (S19).\\\n\\\n**Video:**\\\n\n\n\\\n**Our Ask:**\\\n**If you are a founder or operator at a fintech** **or bank** **who**\\\ni) has a pressing need to automate manual tasks due to high volume and/or,\\\nii) needs to decrease the amount you spend annually on back-office staff or\\\niii) you simply want to build solid automation to decrease human error\\\nReach out directly: [founders@rapidfolio.com](mailto:founders@rapidfolio.com) or book a demo: \n\n**——————-**\n\n**The Problem** \n\nFintechs and banks spend up to **30% of their revenue** on operations, compliance, risk and fraud back-office staff. The back-office performs manual processes like KYC and KYB reviews, OFAC/sanctions/watchlist reviews, fraud investigations, disputes and chargebacks, mortgage and loan underwriting, and sometimes even manual data input workflows, among others. The back-office scales as a function of the number of accounts held at the fintech/bank and quickly becomes extremely costly, time consuming, and error prone. \n\nFrederik and I lived this problem firsthand. **We started Rapidfolio because we wished it existed when we were running Sable (S19)!**\n\n**Our Solution** \n\nRapidfolio turns manual compliance, operations, risk and fraud processes into accurate, auditable, and automated flows. We’re running at **100% accuracy** in production. We provide the **end to end platform and services** to:\n\n**1. Convert your existing procedure** into deterministic, automated, workflows in minutes.\n\n![uploaded image](/media/?type=post&id=106986&key=user_uploads/334623/8df4940f-c6ec-41e1-a6a5-4ccd6476667b)\n\n**2.** We generate dozens of **test cases**, including edge cases, and fully test the workflow.\n\n![uploaded image](/media/?type=post&id=106986&key=user_uploads/334623/940aa8f9-9728-47ef-9a5c-81648c0b28b8)\n\n**3.** Once tested, you can **trigger the workflow** to run it in whichever way you prefer - be it API, our scheduler, a JIRA ticket.\n\n![uploaded image](/media/?type=post&id=106986&key=user_uploads/334623/2ee48bae-3d66-487c-9e70-9ab859eb0b02)\n\n**4. You have full visibility and control:** all workflows are version controlled, and each action is auditable and traceable. You can include humans-in-the-loop where needed. All analytics live directly in your dashboard.\n\nRapidfolio provides “innate abilities” like automated web browsing, document reading/analysis, comprehensive summaries, human-level judgment, to name a few. Additionally, Rapidfolio integrates easily into your **existing tech stack**. We have built-in integrations with leadings operations, compliance and fraud vendors in the financial services industry.\n\nWe are SOC 2 compliant, and are **live with customers across the U.S. and Canada.**\n\n\\\n**The Team:**\n\n**Naveen Qureshi:** CEO and co-founder, Rapidfolio\\\nPreviously was co-founder & CEO of Sable (S19) that I built, scaled and sold. After Sable’s exit, I built and launched additional credit products with [Seen.com](http://Seen.com). Ex Capital One, Amazon, Target. Columbia Business School.\\\n\\\n**Frederik Vanhevel:** CTO and co-founder, Rapidfolio\\\nPreviously built the entire banking stack for two fintechs from scratch. First, as founding engineer at Sable (S19), where he shipped the entire banking stack in a record 3 months, pre-AI. And after Sable’s exit, with Seen.com. Fred also built internal tools for back-office staff at both fintechs, and is passionate about automating the more tedious workflows with Rapidfolio.\\\n\\\nReach out directly: [founders@rapidfolio.com](mailto:founders@rapidfolio.com) or book a demo: \n\n![uploaded image](/media/?type=post&id=106986&key=user_uploads/334623/a7390ea6-c745-4d74-9505-8a42ee86603c)\n\n", "url": "https://www.ycombinator.com/launches/Rpa-rapidfolio-ai-platform-for-compliance-operations-risk-fraud-for-fintechs-and-banks", "votes": 28, "created_at": "2026-07-23T18:42:29.801Z" } ] }, { "id": 31184, "name": "Assemble", "slug": "assemble", "former_names": [ "Serin", "Standard Interaction" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/fdaeafd72d237eaf328842585365c3b7ce49cb7c.png", "website": "https://www.assemble.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Assemble accelerates IT teams by performing development work inside core enterprise systems like ERP, CRM, and HRIS. Our agents help IT teams scope dependencies, build logic, and validate changes across your systems.", "one_liner": "Autonomous IT Delivery", "team_size": 4, "industry": "B2B", "subindustry": "B2B", "launched_at": 1783894724, "tags": [ "SaaS", "Enterprise Software", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/assemble", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/assemble.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/use-assemble", "x_url": "https://x.com/assembleagents", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/fdaeafd72d237eaf328842585365c3b7ce49cb7c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 1578320, "full_name": "Aliyan Ishfaq", "title": "Founder/CEO", "founder_bio": "cs/ai @ stanford | prev. SAIL, LangChain", "is_active": true, "linkedin_url": "https://linkedin.com/in/aliyan-ishfaq", "x_url": "https://x.com/MAliyanIshfaq", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e4fa658850fbf6eb187933c123d915d40d059d30.jpg" }, { "user_id": 1804823, "full_name": "Shaurnav Ghosh", "title": "Founder/CTO", "founder_bio": "Building Assemble.\nSAIL Researcher. Previously at AWS, Apple. Stanford CS BS/MS.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shaurnav-ghosh/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e26c4f24efcca55ac87f67e85a7c991281951379.jpg" }, { "user_id": 3118829, "full_name": "Shrish Janarthanan", "title": "Founder/CPO", "founder_bio": "CS @ Stanford. Cyber Security @ KPMG. Mayfield Fellow. Building AI for IT Teams. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shrishj/", "x_url": "https://x.com/shrishj5", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6a6deb9a94d9a3b0b11ae04c5d045d3c8d4e5888.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 105577, "title": "Assemble: Agents for Enterprise IT", "tagline": "Agents that perform IT implementations across ERPs, CRMs, and HRIS platforms.", "body": "Hi, we’re Assemble!\n\nWe build agents for IT teams that manage enterprise systems like ERP, CRM, and HRIS platforms. Think Salesforce, SAP, Workday, NetSuite, and custom applications. \n\nThese **platforms keep the world’s businesses running every day**, yet they are painful to configure as the business evolves. With custom logic, integrations, and processes built up over years, IT implementations can take months or even years.\n\nAssemble helps IT teams understand these dependencies, make changes, and test the result, all while keeping a clear agent trace. Teams can review the work, approve it, and roll it back when needed. We are building Assemble to help IT teams deliver complex transformation projects with both speed and control.\n\n**Demo**\n\n\n\nThe demo shows Assemble migrating trade-show accounts from HubSpot to Salesforce and setting up APEX automation for hot leads. The agent completes the work end-to-end, creating an audit trail for every action letting the user revert individual changes or restore the Salesforce to an earlier snapshot.\n\n**Why we started Assemble**\n\nWe learned about the problem space after talking to more CIOs and VPs of IT. Companies are already using AI to ask questions about their core business stack. The real challenge is to have safe-execution inside these core systems, letting IT move just as fast as the business.\n\n![uploaded image](/media/?type=post&id=105577&key=user_uploads/1578320/4aa47909-92cb-46a2-9479-917d889f318c)\n\nShrish (left): Cyber Security at KPMG, Mayfield Fellow. \n\nAliyan (center): Researcher at Stanford AI Lab, Applied AI Engineer at Langchain, built products used by 100k+ people\n\nShaurnav (right): AI Engineer at Amazon and Apple. Researcher at Stanford AI Lab\n\nWe have worked on long-running agents and are excited to apply that experience to enterprise IT.\n\n**Asks:**\n\n1. If your company uses Salesforce, NetSuite, Workday, ServiceNow, SAP, or similar systems, we’d love to learn how your IT / business systems team handles changes today. **Are they backlogged? We might be able to help!**\n2. If you know someone who leads IT, Business Systems, or Enterprise Applications at a scaling company, we’d be very grateful for an intro!\n\nTalk to us at [founders@assemble.ai](mailto:founders@assemble.ai)", "url": "https://www.ycombinator.com/launches/RSr-assemble-agents-for-enterprise-it", "votes": 4, "created_at": "2026-07-13T16:23:12.785Z" } ] }, { "id": 31216, "name": "Archal", "slug": "archal", "former_names": [ "Archal Labs" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/d5457b412312bd26ecc72ccead99d16248cd087a.png", "website": "https://www.archal.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Archal provisions sandboxes with the API environments your CI, tests, and evals need. Your coding agent can start the sandbox & environments, connect to your API integrations, run tests, and inspect state changes. Sandboxes and environments are startable within seconds and take milliseconds to reset.", "one_liner": "API sandboxes, built for AI agents", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1776709029, "tags": [ "Developer Tools", "API", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/archal", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/archal.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/archal-labs", "x_url": "https://x.com/archal_labs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/d5457b412312bd26ecc72ccead99d16248cd087a.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 2037605, "full_name": "Aidan Tiruvan", "title": "Founder/CEO", "founder_bio": "Founder at Archal; API sandboxes built for AI agents\nPrev ML research at Scale AI & NASA", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aidantiruvan/", "x_url": "https://x.com/AidanTiruvan", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6d8e1a01ea012584fc41d1ef014c296c0bbdbd00.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111649, "title": "Archal: API sandboxes, built for AI agents", "tagline": "Test API integrations without test accounts, rate limits, or messy state.", "body": "> **TLDR:** Create an API sandbox with the environments your CI, tests, and evals need. Your coding agent can start the API sandboxes (e.g Linear, Slack, Datadog), run tests, and inspect state changes. Ask your agent about archal.ai!\n\nHi everyone, I'm Aidan, the founder of Archal. I previously built agent environments at Scale AI and I've been an avid developer for most of my life, which led me to the problem Archal poses to solve.\n\n[https:/youtu.be/pUfu1RFDYVk](https://youtu.be/pUfu1RFDYVk)\n\n**What we found**\n\nCoding agents are getting much better at writing integrations, but testing those integrations is still painful. Many providers do not offer a usable sandbox, so developers fall back to throwaway accounts, shared test workspaces, or thin mocks. Those options are difficult to reset and make it hard to know whether an agent actually completed a task or only returned a plausible response.\n\nAs Adam Bender at Google put it, “But with 10x more code and 10x more services, integration tests are going to become the most important part of your quality strategy.” (Google, Software engineering is at the tipping point)\n\n**How Archal works**\n\nWith all of this in mind, I built Archal for a starting point of how the next generation of devs will do QA. It's a tool focused on giving agent developers and integration engineers API sandboxes of the SaaS platforms they use in prod pipelines. With Archal, a developer or their coding agent can start a sandbox, run the task, inspect what changed, and reset everything for the next test. We eliminate the need of separate test accounts, thin mocks and stubs, as well as overburdening rate limits on provider-owned sandboxes.\n\n**Our ask**\n\nWe're giving out $20 in Archal credits to each new developer that signs up and starts a sandbox. We’d love for you to give it a try and give us feedback, cheers to making testing easier in the era of coding agents!", "url": "https://www.ycombinator.com/launches/T2n-archal-api-sandboxes-built-for-ai-agents", "votes": 61, "created_at": "2026-08-22T21:00:00.317Z" } ] }, { "id": 31223, "name": "Datoric", "slug": "datoric", "former_names": [ "Flow AI", "Arzule" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/fa894803120df3a3a9250c66e4232e33e2a223d0.png", "website": "https://www.datoric.com", "all_locations": "San Francisco, CA, USA", "long_description": "Datoric develops custom datasets for voice models, robotics, and world models, treating research, collection, verification, and production as one continuous process. We work closely with frontier model teams to turn emerging limitations and model failures into testable data hypotheses, while running our own experiments ahead of customer demand. This allows us to operationalize validated methods into repeatable collection systems at scale.\r\n\r\nData is collected through private invite-only applications separated by modality, customer, and trust level. Every submission remains linked to the contributor, device, task, session, consent, rights, and processing history that produced it, giving our internal QA and fraud models the context to detect problems that may appear legitimate in the finished file. Each collection reveals new failure cases and quality signals that improve the systems behind the next dataset.\r\n\r\nOnce a collection method is validated, we scale it up as a solution to the model failure and it also becomes a reusable data recipe for future custom projects or independently collected, rights-cleared data products. ", "one_liner": "The secure training data R&D engine for physical AI", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1767154035, "tags": [ "Machine Learning", "Robotics", "Data Labeling", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/datoric", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/datoric.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/arzulehq/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/fa894803120df3a3a9250c66e4232e33e2a223d0.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 1486120, "full_name": "Nikhil Reddy", "title": "Founder/CEO", "founder_bio": "CEO @ Datoric. Prev Quant & SWE Intern. Math/Econ/CS @ UChicago; Bypassed Google OAuth 2.1 and QA systems on data annotation platforms to automate training tasks in 2023, then disclosed the loopholes to the affected platforms.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/nikhilrreddy", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/85cf42094155ddf41c9ca0e99b42a85e2ed763d1.jpg" }, { "user_id": 2310490, "full_name": "Jeffrey Lin", "title": "Founder/CTO", "founder_bio": "CTO @ Datoric. Prev AI/ML & SWE Intern. Math & CS & Robotics at NYU. Created a bounding box algorithm for more efficient labeling as Head of Data for the NYU Robotics Team. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jeff-j-lin", "x_url": "https://x.com/Jeffjinlin", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4f3068998c884af4a2668f4da4bbe3efa8e796d3.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108251, "title": "Datoric | Trustworthy data for the next generation of models", "tagline": "Collecting the human experience AI needs, securely.", "body": "Datoric provides custom, security-first training data for voice models, robotics, and world models. Our private, project-specific contributor systems have led to a dramatic improvement in quality for customers and helped us generate nearly seven figures in revenue over the last 30 days.\n\n## **Our background**\n\nIn 2023, we built large-scale automations that completed paid LLM training tasks across major crowd-work platforms by bypassing QA testers. We earned six figures in three months, and even helped fix some of the vulnerabilities that we found. From this experience, we learned just how easily training data can be compromised, and how it can lead to unsafe model behavior.\n\n\n\n## **The problem**\n\nLabs are searching for the high-quality, authentic human data their models need, but quality and provenance become harder to control at scale, especially when relying on outside data providers. Poor-quality or compromised data can distort model behavior, while unqualified contributors, automated submissions, unclear instructions, and weak consent records can still slip through QA. At the same time, data providers hold concentrated stores of customer data, contributor identities, and payments that attract fraud and security threats.\n\n## **What we built**\n\nDatoric manages the full collection process for multilingual speech, egocentric video, and action-conditioned data. We collect this data through more than 300,000 active contributors who record conversations, everyday tasks, and other project-specific activities through our private apps. Because Datoric has no public marketplace or open upload system, every collection can run inside its own isolated environment, keeping customer projects separate and limiting unauthorized access. Each recording also remains linked to the verified contributor who created it and the consent they provided. This blocks automated or manipulated submissions and protects quality throughout the collection process rather than leaving it to final QA.\n\n## **Traction**\n\nWe’ve seen a major uptick in quality compared to other providers, and in the last 30 days, Datoric has generated nearly seven figures in revenue.\n\n## **Our ask**\n\nWe would love introductions to teams building voice models, robots, or world models that need difficult, custom human-generated data.\n\nSend us your hardest collection specification at [**founders@datoric.com**](mailto:founders@datoric.com), or visit [**datoric.com**](http://datoric.com).", "url": "https://www.ycombinator.com/launches/S9z-datoric-trustworthy-data-for-the-next-generation-of-models", "votes": 4, "created_at": "2026-07-29T16:22:25.416Z" } ] }, { "id": 31226, "name": "Libra Robotics", "slug": "libra-robotics", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/e588f98ac537da3d55ea3c2b43b0bd505524c7f8.png", "website": "https://www.librabots.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Infrastructure faces rapid expansion and labor shortages. Today, robots struggle to scale to meet customers’ high-mix work needs. Libra builds AI that engineers robots for new tasks across embodiments. We go from each new infra task to field-deployable robots 10x faster. Today, our robots are working on solar farms and piloting in data centers. Contractors pay for completed work as a service, not robots.", "one_liner": "The robot workforce for infrastructure", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1783647531, "tags": [ "Hard Tech", "Robotics", "Construction", "Energy", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/libra-robotics", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/libra-robotics.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/librabots", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/e588f98ac537da3d55ea3c2b43b0bd505524c7f8.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 3178817, "full_name": "Jermaine Zhao", "title": "Founder/CEO", "founder_bio": "Full-stack robotics engineer with an EE & ME background from Stanford. Built and deployed robots across industrial automation, agriculture, and utility-scale solar. Founder of Libra Robotics.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jermaine-zhao-9439202a2/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/54d431097cd70fba6fbe8902e21879a923482e82.jpg" }, { "user_id": 2782328, "full_name": "Yaojing Huang", "title": "Founder/COO", "founder_bio": "Stanford Civil Engineering+CS | PhD defended in Construction Robotics\n\nDissertation on shared robotic platforms and reusable construction skills across tasks. Research at Center for Integrated Facility Engineering and the Stanford Robotics Center; taught Stanford’s Construction Robotics course; worked with 40+ contractors and robotics companies worldwide.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/yaojing-h-0b903620b/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9f3184f01d6822e5211831c041bd669f65ee31b5.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 122276, "title": "Libra Robotics: The robot workforce for infrastructure", "tagline": "One intelligence layer, many robots. Across solar farms, data centers, and beyond.", "body": "Hey everyone 👋 We're Jermaine and Yaojing, co-founders of Libra Robotics. **We'll present & demo at YC's F26 Demo Day!** If you're attending, we'd love to connect.\n\n**TL;DR:** Infrastructure faces rapid expansion and labor shortages. Today, robots struggle to scale to meet customers’ high-mix work needs. Libra builds AI that engineers robots for new tasks across embodiments. We go from each new infra task to field-deployable robots 10x faster. Today, our robots are working on solar farms and piloting in data centers. Contractors pay for completed work as a service, not robots.\n\n\n\n\\[A simplified walkthrough of how Libra’s AI takes a real data center customer's task from initial requirements to robotic deployment - autonomously designing, simulating, testing, and improving along the way.\\]\n\n**The Problem**\n\nInfrastructure has a labor shortage. Robotics has a scalability problem.\n\nUnlike factories, infrastructure work is high-mix: tasks, designs, and environments constantly change. Single-task robots need custom engineering per task; general-purpose robots need enormous data to become reliable. **Neither scales to thousands of changing tasks.**\n\n**Our Approach**\n\nInstead of building one robot that does everything, we build **reusable robotic capabilities and AI that engineers them into new tasks.**\n\nMany infrastructure tasks share the same physical challenges: handling heavy components, precise positioning, alignment, and fastening. We've built a common set of robotic capabilities around these operations, designed to work across different robots and installation tasks.\n\nOur hardware-agnostic AI selects the right robot configuration and tooling, reuses existing skills, and engineers what is missing for each new task, reducing the need for custom engineering.\n\nIt validates solutions in simulation, then autonomously designs and runs physical experiments on our in-house robotic testbeds, stress-testing solutions against real-world variability (such as unexpected obstacles, changing environmental conditions, etc.). When a test fails, the system **diagnoses the failure, recovers, and iterates autonomously** - without human engineers in the loop. Through **recursive self-improvement (RSI)**, it continuously refines its approach toward field-ready performance.\n\n**Every task adds reusable knowledge, making the next one faster and cheaper to automate.**\n\n**GTM & Traction**\n\n**We deploy AI-powered robot crews to perform infrastructure work**, starting with solar installation and data center electrical work, including prefab MEP and busway installation. \n\n**We're already beyond the lab:**\n\n→ **Robots already deployed** on real solar construction sites\n\n→ **Signed commercial agreements** for solar installation\n\n→ **Data center pilots underway** for MEP automation\n\nWe land with one high-value task, then expand into more work for the same customer. Our goal is to achieve 100 million hours of robotic work annually by 2031.\n\n![uploaded image](/media/?type=post&id=122276&key=user_uploads/3178817/7e72e897-d6d4-4ed6-a547-0f5889a68de0)\n\n\\[Libra’s robots in action on solar farms\\]\n\n**Who We Are**\n\nYaojing did her Stanford PhD on the exact problem Libra solves: one shared robotic platform with reusable skills that generalizes across construction tasks. She has worked with 40+ contractors and construction robotics companies.\n\nJermaine previously led a team to build and deploy autonomous agricultural harvesting robots from the ground up, tackling complex manipulation, precision harvesting, and multi-robot coordination in unstructured real-world environments. He brings end-to-end experience turning robotics research into reliable, field-deployed systems.\n\n**Our Ask**\n\nWe're looking for introductions to:\n\n* **Data center GCs, EPCs & Prefab contractors** interested in deploying robots for prefab MEP installation, busway installation, and other electrical fit-out tasks.\n* **Solar EPCs & installers** with upcoming utility-scale projects interested in robotic installation.\n* **Robotics engineers & researchers** working on sim-to-real, autonomous robot engineering, and multi-embodiment learning who want to build with us.\n* **Investors** who are interested in backing the robot workforce companies for AI era.\n\nKnow someone we should talk to? We'd love an introduction!\n\nReach us at [founders@librabots.com](mailto:founders@librabots.com).", "url": "https://www.ycombinator.com/launches/VoC-libra-robotics-the-robot-workforce-for-infrastructure", "votes": 14, "created_at": "2026-10-15T15:49:58.924Z" } ] }, { "id": 31234, "name": "Ekpa", "slug": "ekpa", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/06c4988b8eff9db30a66dc2c4f5f513eba67b606.png", "website": "https://goekpa.com/", "all_locations": "Boston, MA, USA", "long_description": "", "one_liner": "Building Autonomous Research Agents for Trading", "team_size": 2, "industry": "Fintech", "subindustry": "Fintech -> Asset Management", "launched_at": 1783109218, "tags": [ "Consumer Finance" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech", "Asset Management" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/ekpa", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/ekpa.json", "year_founded": 2026, "city": "Boston", "country": "US", "linkedin_url": "https://www.linkedin.com/company/ekpa/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/06c4988b8eff9db30a66dc2c4f5f513eba67b606.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 3028173, "full_name": "Anant Asthana", "title": "Founder", "founder_bio": "Co-founder @ Ekpa", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/asthana-anant/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/35d5061a5dc077f08c538bfd81cf5371b0b5252a.jpg" }, { "user_id": 2421466, "full_name": "Yuga Patel", "title": "Founder", "founder_bio": "Co-Founder @ Ekpa | CS/AI @ MIT", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/yugapatel26", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2253a03131fee0fb9263a61fbc45c54c1e30c348.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 105026, "title": "Ekpa: AI-Native Investment Platform for Everyday Traders", "tagline": "Intelligent and Data-Backed Investment Decisions", "body": "Hi all, \n\nWe’re Yuga and Anant. Yuga is a CS & Econ major at MIT with investment experience; Anant is CS & ChemE at MIT who has built agentic pipelines for drug discovery.\n\n**TL;DR:** With Ekpa, retail investors can create and improve their portfolio with tailored and strategic advice. Reach out if you are/know anyone who is into investing!\n\n**90 sec demo 👇**\n\n\n\n❌ The problem\n\nThere is a big gap between passive index-fund investors and wealthy people who have access to real hedge funds. Hedge funds have an army of people doing research and tracking and tracing every single piece of data. \n\nA retail investor with a full-time job and family does not have time to do all the research and make intelligent trading or investing decisions. \n\nAn everyday person lacks time, tools, expertise, and therefore confidence to make decisions that bring them closer to achieving financial freedom. \n\nTherefore, they miss out on the opportunity to build their wealth unlike the wealthy who continue to compound their wealth. \n\n**✅ The solution**\n\nWhat Ekpa enables:\n\n* **AI-Native Data Workflows** - agents that constantly keep up with the market trends, news, SEC filings, 10k, etc based on your portfolio.\n* **Research Systems**, based on your current portfolio and current interest. \n* **Discover New Investments** based on your risk-tolerance and future goals.\n* **Active Alert System**, so you don’t miss out on when the markets are moving while spending time doing the things you love. \n\nWith Ekpa, individuals build wealth regardless of experience. We’ve built AI trading systems that have outperformed the S&P500 over short time periods (over +1.5% returns in one week). \n\n**The ask**\n\nWe’re onboarding early customers and would love intros to anyone interested in trading\n\nFeel free to reach out directly to us at founders@goekpa.com. ", "url": "https://www.ycombinator.com/launches/RJy-ekpa-ai-native-investment-platform-for-everyday-traders", "votes": 17, "created_at": "2026-07-09T15:29:01.604Z" } ] }, { "id": 31249, "name": "Meteoric", "slug": "meteoric", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/d2429e5b744af3176c949e0f696004529640d5c5.png", "website": "https://meteoric.earth/", "all_locations": "San Francisco, CA, USA", "long_description": "Meteoric is building drones that clear clouds over solar farms and ultimately weaken severe storms and hurricanes.", "one_liner": "Drones that clear clouds over solar farms and weaken hurricanes", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Climate", "launched_at": 1786601879, "tags": [ "Hard Tech", "Renewable Energy", "ClimateTech" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Climate" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/meteoric", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/meteoric.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://uk.linkedin.com/company/meteoric-inc", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/d2429e5b744af3176c949e0f696004529640d5c5.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 2005935, "full_name": "Mete Karslioglu", "title": "Founder/CEO", "founder_bio": "Building drones that clear clouds over solar farms and ultimately weaken severe storms and hurricanes. \nPreviously developed spray nozzles to brighten clouds (MCB) at the Cambridge Centre for Climate Repair. \nCambridge MEng in Aerospace, Energy & Mechanical Engineering \n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/mete-karslioglu/", "x_url": "https://x.com/MeteKarsli0glu", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b998b2251a59ddfe79caef102ec082aabdfe46ae.jpg" }, { "user_id": 2007524, "full_name": "Eric Nilsson", "title": "Founder/CTO", "founder_bio": "Co-founder and CTO at Meteoric (S26). We build drones that clear clouds over solar farms and ultimately weaken hurricanes. \nPreviously worked on capacitive sensing for digital microfluidics at the University of Cambridge.\nCambridge MEng in Engineering, specialised in Electrical & Electronic Engineering, Information & Computer Engineering, Instrumentation & Control.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/eric-nilsson-2b3775254/", "x_url": "https://x.com/eric_nilsson_", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9396f9942c88c3920b9d6d17d32c28275e141a56.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111583, "title": "Meteoric — Drones that clear clouds over solar farms and ultimately weaken hurricanes", "tagline": "Meteoric is building drones that reduce the reflectivity of clouds above solar farms and ultimately reduce the intensity of severe storms and hurricanes.", "body": "![uploaded image](/media/?type=post&id=111583&key=user_uploads/2005935/6f49c082-1391-4bb6-939f-8f92d0dee7a0)\n\n**TL;DR:** Meteoric is building drones that fly into clouds and reduce their reflectivity without spraying any chemicals. We're starting with clearing overcast clouds above utility-scale solar farms to increase their annual generation by 10–30% with no new infrastructure. Our long-term goal is to scale this system to weaken severe storms and hurricanes. First large-scale cloud clearing flight in 2027 and first storm flight in 2028.\n\nHey everyone! We’re [Mete](https://www.linkedin.com/in/mete-karslioglu) and [Eric](https://www.linkedin.com/in/eric-nilsson-2b3775254), and we’re building drones to engineer clouds.\n\n![uploaded image](/media/?type=post&id=111583&key=user_uploads/2005935/5c79de40-e828-4e83-aff5-e1ae0feb3dc4)\n\n⚡ **Starting with solar**\n\nAI has created an enormous demand for power, and solar is scaling exponentially. [US utility-scale solar capacity is set to double by 2030](https://cleanpower.org/resources/market-pulse-utility-scale-solar-summer-2026), and [the world will add 2+ TW](https://www.iea.org/reports/renewables-2025/renewable-electricity) over the same period.\n\nBut building new capacity is getting harder and less profitable. Only [\\~14% of US utility-scale solar projects](https://emp.lbl.gov/queues) actually get built and [773 GW is already waiting in interconnection queues](https://emp.lbl.gov/queues). \n\nAt the same time, generation costs are rising as tax credits are overhauled and tariffs increase material prices. This makes each additional MWh generated from an existing plant more valuable.\n\n[Low and mid-level overcast clouds cut incoming sunlight by 73–82%](https://www.sciencedirect.com/science/article/pii/0038092X80903916) while they’re overhead. We recover that without building new infrastructure. \n\n🌤 **What we're building**\n\nFleets of 10s-100s of autonomous drones that fly into low and mid-altitude clouds (1–5 km) over a solar farm and alter the water droplets that form them without using any chemicals. This reduces their reflectivity and brings back measurable amounts of sunlight to the panels below.\n\nOur cloud-loss model shows that this process can increase annual generation from existing assets by 10–30% across major US grid regions, worth \\~$5,000–28,000/MW, without building any new infrastructure.\n\n⚙️ **Traction**\n\n* Built a working prototype that dissipated an artificial cloud by 13% in cloud-chamber tests\n* Built a cloud-loss model that quantifies how much generation a given US solar plant is losing to the clouds we can treat\n\n👉 **Why now**\n\n* Cloud access got \\~97% cheaper: conventional seeding aircrafts cost $2,000+/hr; electric drones are at $30–60/hr. \n* AI is transforming weather forecasting: DeepMind's weather models produce [15-day forecasts in less than a minute](https://deepmind.google/blog/weathernext-ai-model-achieves-breakthrough-in-forecasting-cyclones/) and [outperform traditional physics-based models on more than 90% of benchmarks](https://developers.google.com/weathernext/guides/evals).\n\n🎯 **Weakening storms and hurricanes**\n\nOur ultimate goal is to reduce the intensity of severe storms and hurricanes.\n\nSevere weather has already cost the US [$31.9B in the first half of 2026](https://www.climatecentral.org/climate-services/billion-dollar-disasters), and from 2020 to 2024 severe storms, floods, and tropical cyclones caused [an average of \\~$121B/yr in losses](https://www.ncei.noaa.gov/access/billions/summary-stats). Severe storms alone generated [$50B+/yr in insured losses for the past three years](https://www.iii.org/press-release/triple-i-severe-convective-storms-generate-more-than-50b-in-insured-losses-for-third-consecutive-year-041326). There's now [a billion-dollar weather disaster every 16 days vs. every 82 days in the 1980s](https://www.ncei.noaa.gov/access/billions/climatology).\n\nGlobally, these disasters caused [over $2T in damage in the past decade](https://iccwbo.org/news-publications/policies-reports/new-report-extreme-weather-events-cost-economy-2-trillion-over-the-last-decade) and claimed [more than 2 million lives in the past 50 years](https://news.un.org/en/story/2023/05/1136897). \n\nClearing stable overcast clouds is our first major technical milestone. It will prove we can reliably alter clouds at scale. From there, we’ll expand our fleets and control systems to reduce the intensity of severe storms and hurricanes. First storm operation targeted for late 2028.\n\n👥 **Team**\n\nWe're two Cambridge engineers with backgrounds in weather engineering and droplet-control systems:\n\nMete (CEO): Master's research on making clouds more reflective using seawater spray nozzles at the Cambridge Centre for Climate Repair (funded by ARIA, UK's DARPA). MEng from Cambridge in aerospace/aerothermal, energy, and mechanical engineering.\n\nEric (CTO): Master's research on microfluidic chips that control water, and built droplet-removal technology for LiDAR lenses. MEng from Cambridge in electrical, control, and computer engineering. Placed 4th nationally in the British Physics Challenge.\n\n🙏 **We'd love to hear from you if you:**\n\n* work in disaster response or severe weather at the state or federal level (FEMA, state EMs, NOAA, NWS)\n* run hail suppression programs (North Dakota, Texas, Alberta)\n* own, manage, build, or operate utility-scale solar plants (especially in Texas, northern US, UK, or northern Europe)\n* work in weather modelling / meteorology / atmospheric science (especially in hurricane or storm research)\n* are engineers with experience in drones or high-voltage systems / power electronics\n\nReach us at [team@meteoric.earth](mailto:team@meteoric.earth)\n\n![uploaded image](/media/?type=post&id=111583&key=user_uploads/2005935/c0a9ce1e-492d-4687-8c6d-4d53a7bf67b2)\n\n", "url": "https://www.ycombinator.com/launches/T1j-meteoric-drones-that-clear-clouds-over-solar-farms-and-ultimately-weaken-hurricanes", "votes": 19, "created_at": "2026-08-21T20:00:00.552Z" } ] }, { "id": 31257, "name": "Atlia", "slug": "atlia", "former_names": [ "Odyssey" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/7a4b2fb5aaa99f5f795d2aa898508681eae26222.png", "website": "https://www.atlia.com", "all_locations": "San Francisco, CA, USA", "long_description": "We use AI agents to manage short-term rental properties by running operations completely autonomously. We cut traditional human management fees in half so owners take home 20% to 30% more revenue. And we operate with near-zero boots-on-the-ground so we can scale nationwide with software margins and at venture speed in a traditionally physical industry.", "one_liner": "The AI-native property management company for short-term rentals", "team_size": 2, "industry": "Real Estate and Construction", "subindustry": "Real Estate and Construction -> Housing and Real Estate", "launched_at": 1785302953, "tags": [ "Real Estate", "Proptech", "Big Data", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Real Estate and Construction", "Housing and Real Estate" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/atlia", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/atlia.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/atlia", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/7a4b2fb5aaa99f5f795d2aa898508681eae26222.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1330444, "full_name": "Edmond Niu", "title": "Founder", "founder_bio": "Co-founder @ Atlia, enabling properties to be self-managing - continuously learning and improving themselves.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/edmond-niu/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/eefe6358ce17774a69778c1a39a2e87514240345.jpg" }, { "user_id": 1730500, "full_name": "Shaan Yadav", "title": "Founder", "founder_bio": "Co-founder at Atlia, systems for properties to operate autonomously.\nPrev. gpus at Apple, brain-computer interfaces at Yale and computer architecture research at Duke.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shaanyadav", "x_url": "https://x.com/TheShaan", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b80a700259f368cf6a3acdebd10f537a89c3427c.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108283, "title": "Atlia: The AI-Native STR Property Management Company", "tagline": "The AI-Native STR Property Management Company", "body": "### TL;DR\n\n**Atlia is the AI-native short-term rental property management company.**\n\n**You give us your Airbnbs to manage, and you’ll take home 10-20% more while both you and your guests have a better experience.**\n\nLaunch Video:\\\n\n\n## The Problem\n\nShort-term rentals were sold as passive income. But in reality, every Airbnb is a tiny hotel and a nightmare of operations. Every reservation restarts the operational clock: prices must be adjusted, guests screened, messages answered, cleaners scheduled, access coordinated, supplies restocked, and maintenance handled, often overnight or between back-to-back stays.\n\nAnd in traditional short-term property management companies, operational failures are common, and they can quickly become very costly:\n\n* A missed cleaning becomes a bad review\n* A poorly screened guest becomes property damage\n* A neglected maintenance issue becomes an issued refund or lost booking\n\nOn top of these operational failures, the average management company takes 15–30% of rental revenue, often alongside cleaning, maintenance, and other fees.\n\nWe’ve been property managers for the past year and ran into the same lose-lose choice every owner faces: manage a 24/7 hospitality business yourself, or hand over a quarter of your revenue to a management company whose costs increase with every property it adds.\n\n## The Solution\n\nWe are building a new type of property management company, with infrastructure built for agents and human-computer-interaction from the ground up. By replacing the layers of coordination and operational overhead built into traditional management companies, we provide a much better service at a much lower fee.\n\nWe center around a simple premise: owners shouldn’t have to choose between good service and reasonable fees. With our model, **owners take home 10–20% more and guests have a more premium experience.**\n\n## Team\n\nWe’re Shaan and Edmond, longtime friends, Duke roommates, and property managers.\n\nEdmond previously built AI voice agents at Rilla, LLM agents at S&P Global, and data pipelines at Vanguard. He has been managing rentals for a year, experiencing the owner’s dilemma firsthand.\n\nShaan previously designed GPU hardware at Apple, brain-computer interface research at Yale, computer architecture research at Duke. He also has been managing Airbnbs and longer-term rentals for a year and a half.\n\n## The Ask\n\nWe are live in 80 properties today, and would love to have you join us.\n\n**If you own a short-term rental** (such as an Airbnb), we can help you generate more revenue while providing better experiences to guests and giving you that peace of mind that your property is in good hands.\n\n**If you are considering investing in a short-term rental,** we can help turn your investment into a much more lucrative investment vehicle.\n\nYou can reach out to us at [founders@atlia.com](mailto:founders@atlia.com). And schedule a free property review at [www.atlia.com](http://www.atlia.com).", "url": "https://www.ycombinator.com/launches/SAV-atlia-the-ai-native-str-property-management-company", "votes": 11, "created_at": "2026-07-29T17:19:12.682Z" } ] }, { "id": 31258, "name": "Dream", "slug": "dream", "former_names": [ "Dream Browser" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/479d7e4f9e6b74b1df9f87c67fe614babba0a9b5.png", "website": "https://www.pingdream.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Dream makes AI cameras that catch damage teams miss. Car dealerships, equipment rental companies, fleets, and more move expensive assets in and out all day, and undocumented damage is a costly leak.\r\n\r\nDream cameras are simple to set up and work automatically. Each vehicle or machine that passes is photographed from every angle, identified, and reconstructed into a complete condition record. On return or intake, Dream instantly flags new dents, scrapes, and changes your team might miss, giving one source of truth for every asset's condition over time.\r\n\r\nRyland worked at X/xAI and built aerial damage-detection vision at Vexcel Imaging. Alexander worked at AWS and trained edge-camera models deployed in the wild across Africa and India to detect poachers.", "one_liner": "Using VLMs to Catch Damage", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1786646316, "tags": [ "Hardware", "Machine Learning", "Computer Vision", "B2B", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/dream", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/dream.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/pingdream", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/479d7e4f9e6b74b1df9f87c67fe614babba0a9b5.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1393934, "full_name": "Ryland Birchmeier", "title": "Founder", "founder_bio": "Ryland Birchmeier is Co-founder of Dream (YC S26), building pocket-sized, battery-powered AI cameras that automatically capture assets and catch damage teams miss at car dealerships, equipment rental companies, and more.\n\nPreviously, Ryland worked at X/xAI and built aerial computer vision at Vexcel Imaging. He studied computer science at the University of Virginia.", "is_active": true, "linkedin_url": "https://linkedin.com/in/rylandbirchmeier", "x_url": "https://x.com/rylandbirch", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/37f594b7f6b8754e9966db95609cba5942fd3185.jpg" }, { "user_id": 1456894, "full_name": "Alexander Halpern", "title": "Founder", "founder_bio": "Alexander Halpern is Co-founder of Dream (YC S26), building pocket-sized, battery-powered AI cameras that automatically capture assets and catch damage teams miss at car dealerships, equipment rental companies, and more.\n\nPreviously, Alexander built agents for AWS Marketplace and trained models for edge computing cameras deployed across Africa and India to detect poachers hunting endangered wild cats at Panthera. He studied computer science at the University of Virginia.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/alexanderhalpern", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4702d251513a653e5c3f7e4d1307df73e45916e3.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110473, "title": "Dream - Pocket-sized AI cameras that catch asset damage", "tagline": "Dream makes pocket-sized, battery-powered AI cameras that catch damage teams miss.", "body": "Hey everyone 👋, we are Ryland and Alexander, co-founders of Dream.\n\n**TL;DR: Dream automatically photographs every vehicle or machine that passes through your gate, builds a 3D condition record, and uses AI to catch new damage the moment an asset passes through.**\n\n**Launch Video: [https://youtu.be/uMP12lCb12E](https://youtu.be/uMP12lCb12E)**\n\n**The problem**\n\nDamage is an expensive leak for anyone who moves large, expensive assets in and out all day. It shows up two different ways:\n\n**Car dealership service departments**\n\n* As a car pulls into the service entrance, damage that could be identified and fixed goes unnoticed, which is lost revenue.\n* Without a record of the car's condition on arrival, disputes come down to the customer's word against the dealership's, and the dealership gets blamed for pre-existing damage.\n\n**High-value asset rentals (heavy equipment, trucks, boats, planes, trailers)**\n\n* Teams are inconsistent about photographing assets as they go out and come back, so new damage gets missed before the customer is billed.\n* When damage is missed, there's no evidence to charge the customer, so the company eats the repair cost out of pocket.\n\n**What Dream does**\n\nDream cameras sit at your gate, lane, or service drive and turn every pass into evidence, automatically:\n\n* **Continuous capture:** every vehicle or machine that moves through is photographed from every angle. No one has to remember to do it.\n* **Asset identification:** Dream knows exactly what you need to look at.\n* **3D reconstruction:** every pass becomes a complete 3D condition record for that asset.\n* **AI damage detection:** on return or intake, Dream compares against the last record and flags new dents, scrapes, and visible changes right away.\n\nThe result is one clear source of truth for every asset's condition over time: rental companies catch damage before they bill so they stop eating repair costs, and dealership service departments find more fixable damage and resolve customer disputes with evidence.\n\nWorks across cars, vans, trucks, heavy equipment, and more.\n\n**Why us**\n\nWe have spent years training models for aerial imaging and edge computer vision devices. Ryland has worked on the exact problem of detecting damage to assets from aerial imaging, such as degraded paint. Alexander has trained models that run on edge cameras deployed in the wild in Africa and India to detect poachers.\n\n**Our ask**\n\nWe're onboarding partners now. **We'd love intros to:**\n\n* **Car dealerships**\n* **High-value asset rental companies** (heavy equipment, trucks, boats, planes, trailers)\n\nIf you run one of these, or know someone who does, please message us on LinkedIn!", "url": "https://www.ycombinator.com/launches/Sjp-dream-pocket-sized-ai-cameras-that-catch-asset-damage", "votes": 7, "created_at": "2026-08-14T17:00:00.348Z" } ] }, { "id": 31259, "name": "Tsenta", "slug": "tsenta", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c0363e9662e7db43e0292d57332e9650487ebdbb.png", "website": "https://tsenta.com", "all_locations": "San Francisco, CA, USA", "long_description": "Tsenta is an AI career agent that matches you with jobs, applies automatically, and gets you the best offers on the market.\r\n\r\nWe grew to ~$3M Annualized in three months. \r\n\r\nWe built it after manually applying to over 3,000 jobs ourselves, so that candidates would always have someone on their side.\r\n\r\nToday, more than 90,000+ job seekers trust Tsenta to automate their job search.", "one_liner": "AI career agent that finds matching jobs & applies on your behalf", "team_size": 2, "industry": "Consumer", "subindustry": "Consumer -> Job and Career Services", "launched_at": 1769729252, "tags": [ "Marketplace", "Consumer", "Recruiting", "HR Tech", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Consumer", "Job and Career Services" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": [ { "question": "How did your company get started? (i.e., How did the founders meet? How did you come up with the idea? How did you decide to be a founder?)", "answer": "We're both international students who applied to 3,000+ jobs between us last year. After spending several hours a day retyping the same fields into Workday, we started building Tsenta for ourselves, all throughout our summer Internships, and launched out of our dorm at Rose-Hulman." }, { "question": "What is the core problem you are solving? Why is this a big problem? What made you decide to work on it?", "answer": "The average student applies to 200-400 jobs per cycle and spends 15-30 minutes per application on menial data entry - hundreds of hours wasted on broken forms instead of networking or preparing for interviews. \r\n\r\nEvery existing tool is either cloud-based and a black-box, or a slow chrome extension - and both are priced unnafordably for students, which is why we built transparent, on-device automation that actually works - and can fill out forms end-to-end." }, { "question": "What is your long-term vision? If you truly succeed, what will be different about the world?", "answer": "To make hiring effortless for both job-seekers and recruiters alike." } ], "url": "https://www.ycombinator.com/companies/tsenta", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/tsenta.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/tsenta/", "x_url": "https://x.com/TsentaAi", "fb_url": "https://www.facebook.com/people/Tsenta/61589214140559/", "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c0363e9662e7db43e0292d57332e9650487ebdbb.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/19b61e161111a48627d2b05fa7b775d64b2001db.png", "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/87b7933b17546e93c0f502654a802b7fa2f59591.png" ], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 2788219, "full_name": "Agnay Srivastava", "title": "Founder/CEO", "founder_bio": "Building the job automation platform I wish I had at 19. Founder @ Tsenta\nCS + AI @ Rose-Hulman | ex-Systems @ VIAVI, MLOps @ GrapheneAI, SWE @ Paytm", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/agnaysrivastava?utm_source=share&utm_campaign=share_via&utm_content=profile&utm_medium=android_app", "x_url": "https://x.com/AgnaySrivastava", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b4c4591c4f93e0c99ecac11cc5810808693feadf.jpg" }, { "user_id": 2538831, "full_name": "Pulkit Gupta", "title": "Founder/CTO", "founder_bio": "Building the job platform I was too busy applying to jobs to make. Founder @ Tsenta.\nCS @ Rose-Hulman | ex-Full Stack @ RelateXR, Dexter Capital, Multibhashi (1M+ downloads)", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/pulkitgupta712/", "x_url": "https://x.com/pulkit_gupta2", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4b167d8bb1cb9eab727ad71d240dc730ea584232.jpg" } ], "jobs": [ { "id": 108069, "title": "Founding Engineer", "url": "https://www.ycombinator.com/companies/tsenta/jobs/hYAeFGp-founding-engineer", "location": "Bengaluru, KA, IN / Bengaluru, Karnataka, IN", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "₹7.5M - ₹15M INR", "equity_range": "", "experience": "3+ years", "visa": "US citizenship/visa not required", "skills": [] } ], "news": [ { "title": " Two Indian students build AI startup 'Tsenta' that automates job applications, bags ₹5 Cr from Y Combinator", "url": "https://startuppedia.in/young-entrepreneur/young-entrepreneurs-founded-tsenta-bags-5-cr-from-y-combinator-12117787", "date": "Jun 30, 2026" }, { "title": "‘Applied to 3,000 jobs, got 10 interviews’: Indian founders build AI to fix hiring, raise Rs 5 crore", "url": "https://www.moneycontrol.com/news/trends/applied-to-3-000-jobs-got-10-interviews-indian-founders-build-ai-to-fix-hiring-raise-rs-5-crore-13961381.html", "date": "Jun 29, 2026" }, { "title": "Tsenta wants to make job applications faster, smarter, and a lot less manual\r\n | YourStory", "url": "https://yourstory.com/2026/06/tsenta-wants-to-make-job-applications-faster-smarter-less-manual", "date": "Jun 11, 2026" }, { "title": "Two local college students are changing the way we apply for jobs - AOL", "url": "https://www.aol.com/articles/two-local-college-students-changing-223748000.html", "date": "Apr 24, 2026" }, { "title": "Rose-Hulman students selected for Silicon Valley startup accelerator – Inside INdiana Business", "url": "https://www.insideindianabusiness.com/articles/rose-hulman-students-selected-for-silicon-valley-startup-accelerator", "date": "Apr 24, 2026" } ], "launches": [ { "id": 108345, "title": "Tsenta: Your AI career agent that applies to jobs & lands you interviews", "tagline": "Tsenta matches you with jobs, tailors your resume & applies on your behalf. We applied to 3,000 jobs by hand and built this so that you don't have to :)", "body": "Hey everyone, Pulkit and Agnay here!\n\n\n\n**TLDR:**\n\nTsenta **matches** you with jobs, **rewrites** your **resume**, and **applies** for you. We built this because we applied to over 3,000 jobs as international students, and no other tool worked for us. 45,000+ people use it now, and over 70% of our paid users have landed interviews.\n\nSign up at: | 25 free applications on us, no card required :) + a 25% launch discount on any plan!\n\n**Why We Built Tsenta:**\n\nBecause job hunting sucks, and existing platforms suck harder. They either don't submit applications end-to-end reliably, don't give you visibility into what's being filled, or cost more per month than most students can afford. Which is why we built Tsenta to be the best platform on the market for job seekers, both feature and price wise.\n\n**What Tsenta does:**\n\n1. Matches you with over two million jobs in our database, updated every couple of hours.\n2. Rewrites your resume and cover letter in ATS friendly formats.\n3. Actually submits job forms on your behalf, handling account creation, email verification, & filling in OTPs on over 30 ATSs.\n4. Tracks all of your applications for you, so you never have to rely on a spreadsheet again.\n\nTsenta also runs on iOS, Android, iMessage, WhatsApp & as a chrome extension in addition to the web, so you can apply on the go, chat with Tsenta to manage and submit jobs, and use the Chrome extension to bring your own jobs as you browse other websites.\n\nSome of the places people have gotten interviews through Tsenta: Cisco, Goldman Sachs, JPMorgan Chase, Intel, NVIDIA, Scale AI, Northrop Grumman, Raytheon, SpaceX and Teamworks.\n\n**Who we are:**\n\nWe first met at a friend’s dinner back in 2024 and have been close friends for over 2 years now.\n\n**Pulkit:** Coding since the age of 13,Built a language app with 1M+ downloads at 17, shipped Bhasha Sangam for the Govt. of India (500k+ users), dealflow tooling at Dexter Capital. CS @ Rose-Hulman '28\n\n**Agnay:** Coding since the age of 11, Open-sourced Android device trees used by 2,000+ ROM builders, systems engineering at VIAVI, CS + AI @ Rose-Hulman '26.\n\n**Our asks**\n\n1. If you are job hunting, or know someone that is, send them to [tsenta.com/r/ycombinator](https://tsenta.com/r/ycombinator). 25% off any plan, first 25 applications free.\n2. We want this to be the best career agent out there and are super product-focused. Tell us what's broken or shoot us ideas at: [founders@tsenta.com](mailto:founders@tsenta.com).\n3. Hiring? We're building the employer side next and would love to hear where your process today breaks. Same email as above :)", "url": "https://www.ycombinator.com/launches/SBV-tsenta-your-ai-career-agent-that-applies-to-jobs-lands-you-interviews", "votes": 60, "created_at": "2026-07-31T16:26:27.894Z" } ] }, { "id": 31261, "name": "Control Seat", "slug": "control-seat", "former_names": [ "C42" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c9fbf6b23216653c383d7f0caea2def1b673a3a5.png", "website": "https://www.controlseat.com", "all_locations": "San Francisco, CA, USA", "long_description": "Control Seat builds the hardware and software that helps industrial teams see, understand, and run their operations.\r\n\r\nOur platform creates one clear source of truth across the plant. It connects the systems teams already use, brings control, monitoring, history, and machine health together, and adds an intelligence layer that continuously watches for meaningful changes. Teams can understand what happened, catch problems earlier, and act before they become downtime.\r\n\r\nThe result is fewer surprises, faster decisions, and less time spent piecing together information across disconnected tools.\r\n\r\nWe work with manufacturers, utilities, energy companies, and other industrial operators modernizing existing facilities or building new ones.", "one_liner": "AI that runs factories. ", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1777448165, "tags": [ "Workflow Automation", "Monitoring", "Industrial" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/control-seat", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/control-seat.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/control-seat", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c9fbf6b23216653c383d7f0caea2def1b673a3a5.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3064044, "full_name": "Warren Shepard", "title": "Founder/CEO", "founder_bio": "Building a modern SCADA system to control and monitor factory hardware and replace legacy systems. Previously at Dartmouth ('26), SpaceX (Starlink), and Microsoft (security).", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/warren-shepard-493a37201/", "x_url": "https://x.com/warrren_shepard", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ff41fa222f3fc3b27bd9e580e1dee71f5754722d.jpg" }, { "user_id": 3132492, "full_name": "Jack Grodnick", "title": "Founder/CTO", "founder_bio": "Building Control Seat, the AI-native controls and monitoring stack for industry. We help industrial teams connect plant systems, understand what changed, catch problems earlier, and run operations with fewer surprises. On leave from Dartmouth. Previously at Tesla, SpaceX, and Amazon.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jack-grodnick/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b49172bc391f2b19dcde4aff68ff061ccd456a44.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 104972, "title": "Control Seat: the operating system for the physical world.", "tagline": "An AI-native controls (SCADA) platform and integration service.", "body": "\\\n\\\n**THE PROBLEM**\n\nIndustrial control software still looks and feels like it was built in the Windows 95 era. Manufacturers spend weeks stitching together Programable Logic Controllers (PLCs), Supervisory Control and Data Acquisition (SCADA), HMIs, databases, and custom integrations just to get systems running. Additionally, legacy systems collect enormous amounts of high-quality data, but don’t do _anything_ with it, even though it could easily be used for things such as anomaly detection or optimization recommendations.\\\n\\\n**OUR SOLUTION**\n\nWe're building an AI-native platform to unify the industrial control stack, starting with SCADA and eventually extending down to PLC hardware and the devices on the factory floor. We offer both a product and a service:\n\n**1) Software**\n\nWe already built a full SCADA system to replace legacy platforms like Ignition, WinCC, and FactoryTalk. It’s AI native: users can use AI to generate controls screens in seconds (this used to take hours), write custom scrips/logical flows, and automatically detect issues before they become catastrophic. \\\n\\\nNot only is this a full SCADA system, but it can operate on top of an existing one so that users can take advantage of our “intelligence layer” without having to rip out their current infrastructure (people understandably don’t like doing this).\\\n\\\n**2) System Integration** \n\nWe’ve recently expanded to become an AI-native industrial integrator. Instead of just selling our software, we integrate it into our clients existing stack for them, and help them set up any other tools they need to run it. \n\nOver the past few months, we've spent hundreds of hours talking with controls engineers, manufacturers, operators, and system integrators. We already knew that industrial software was broken, but we also learned that even when better technology exists, people are understandably cautious about adopting it. These systems run critical infrastructure and reliability matters more than novelty. Change is slow, but we want to reindustrialize America now.\n\n**Instead of waiting for the industry to adopt an entirely new software stack, we're bringing our technology directly into customer projects today.**\n\nWhile larger firms can spend millions on integration services, smaller firms don’t have that kind of money. Traditional integrators use highly manual methods to set up their clients software stacks (which are also outdated). By leveraging AI, we’ll be able to set up the same systems at 10x the speed and charge way less. \n\n**Our ask: if you’re exploring controls/automation systems, PLC programming, SCADA systems, or system integration**, **we’d love to talk. Or, if you know any deep tech companies that want to set up a controls system, let us know!**\n\nSee more at .", "url": "https://www.ycombinator.com/launches/RJ6-control-seat-the-operating-system-for-the-physical-world", "votes": 4, "created_at": "2026-07-09T14:31:33.522Z" } ] }, { "id": 31268, "name": "Cova", "slug": "cova", "former_names": [ "Omnition" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/a88cf729487706988d23e57fa307a54ba446b835.png", "website": "https://cova.care/", "all_locations": "San Francisco, CA, USA", "long_description": "Cova is an AI-native home care agency, licensed in Michigan. We serve Medicaid, LTCI, and private-pay patients; we pay caregivers the highest base rate in the state; and we help unpaid family caregivers get paid through programs they didn't know they qualified for. Our entire operation runs on AI agents, which means less overhead and more value passed directly to the people doing the real work.", "one_liner": "Financial infrastructure for family caregiving", "team_size": 2, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1782162197, "tags": [ "Consumer Health Services", "Healthcare", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/cova", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/cova.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/cova-home-care", "x_url": "", "fb_url": "https://www.facebook.com/profile.php?id=61590134944668", "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/a88cf729487706988d23e57fa307a54ba446b835.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1711673, "full_name": "Jordan Ibe", "title": "Founder", "founder_bio": "Co-founder @ Cova. Prev Palantir, Uber, Cornell CS & Stats. Building infrastructure to support millions of caregiving families.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/c-jordan-ibe/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/92a8316b30c56d164dc6a5236a7ef4134dae6a98.jpg" }, { "user_id": 1452184, "full_name": "Ansh Tandon", "title": "Founder", "founder_bio": "Co-founder @ Cova (YC S26). Prev Palantir, UPenn Bioengineering BSE & MSE. Building infrastructure to support millions of family caregivers.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/anshtandon/", "x_url": "https://x.com/anshtdn", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/aadab04a90be012be8c61f6057caee67519818d2.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109454, "title": "Cova: The home care provider run by AI agents", "tagline": "A home care agency that scales without administrative headcount", "body": "**TL;DR: [Cova](https://cova.care/) is a licensed home care agency in Michigan. AI agents run the entire back office, which lets us pay caregivers the highest base rate in the state, charge families less, and help unpaid family caregivers get paid through programs they didn't know they qualified for.**\n\n\n\n**The problem**\n\nThis is one of the largest, fastest-growing markets in the country. 11,000 Americans turn 65 every day, and 70% want to age at home. 53 million people are already unpaid family caregivers, over $1 trillion in economic value every year that the system doesn’t capture.\n\nBut getting someone into home care is brutal. The application and paperwork alone take weeks of back-and-forth across a caregiver, a patient and their family, a doctor, and a state caseworker, with paper forms bouncing between all of them. Most of that time isn't care; it's coordination. Most families give up before care ever starts, and caregivers wait unpaid.\n\nWhen care finally begins, caregivers suffer from missed payments and poor coordination, affecting patients downstream. All of this because existing agencies aren’t built to handle care at scale.\n\n**What Cova does**\n\nWe own the full stack: Cova is its own home care agency serving Medicaid, LTCI, and private-pay patients, and AI agents run it end to end.\n\n* **Weeks → days.** AI agents handle the applications and paperwork that normally drags on for weeks in just a few days.\n* **Zero FTE overhead = highest caregiver pay in the state.** AI agents do the coordination work, so we carry no non-caregiver headcount and pass that straight to caregivers as the highest base rate in Michigan.\n* **We get family caregivers paid.** Millions are already caring for a loved one for free and qualify for programs they've never heard of. AI agents find those programs and get them enrolled and paid.\n* **We bill the payers directly.** Because we own the agency, we make money the way agencies do, billing Medicaid, MCOs, LTCI, and private pay, just with a fraction of the cost structure.\n\n**About the team**\n\n**Jordan Ibe (Co-Founder)** – Studied Computer Science & Statistics at Cornell University. Previously a Forward Deployed Software Engineer on Palantir’s Healthcare team and Software Engineer at Uber.\n\n**Ansh Tandon (Co-Founder)** – Completed a BSE & MSE in Bioengineering at the University of Pennsylvania. Previously a Deployment Strategist on Palantir’s Healthcare team and Researcher at Daffodil Health.\n\n**Our ask**\n\n* Intros to Michigan MCOs (Medicaid managed care orgs) and long-term care insurance contacts\n* Intros to Michigan hospital systems and providers: Henry Ford Health, Corewell, and similar\n* Intros to Michigan discharge planners, hospital case managers, and social workers who place patients into home care\n* Know a family caring for an aging loved one in Michigan? Send them our way! We'll check for free whether they qualify to get paid:[ cova.care](http://cova.care)\n\nAnsh & Jordan, [founders@cova.care](mailto:founders@cova.care)", "url": "https://www.ycombinator.com/launches/STO-cova-the-home-care-provider-run-by-ai-agents", "votes": 12, "created_at": "2026-08-06T21:19:22.615Z" } ] }, { "id": 31272, "name": "Poth Labs", "slug": "poth-labs", "former_names": [ "Poth" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ee08f2b3965c952b7b91fda669eab1aace70f2b1.png", "website": "https://pothlabs.com", "all_locations": "San Francisco, CA, USA", "long_description": "Poth is an agentic research platform that connects to your company data, creates hypotheses about what might be causing problems, and validates them using both existing data and adaptive user feedback.", "one_liner": "The customer brain that connects data to explain customer behavior", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1782326273, "tags": [ "SaaS", "B2B", "Workflow Automation", "Market Research", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/poth-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/poth-labs.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://linkedin.com/company/poth", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ee08f2b3965c952b7b91fda669eab1aace70f2b1.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 2622339, "full_name": "Mojmir Horvath", "title": "Founder", "founder_bio": "Cofounder & CTO of Poth Labs. Ex-Tietoevry and Phillips Academy", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/mojmir-horvath/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3251bb7283370c7779f42f2224497630d2fc08a7.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 106978, "title": "Ask Poth—ask anything about your customers", "tagline": "The internal voice of the customer", "body": "# \\*\\*\\*\\*\n\n**TL;DR: Your customer data is fragmented across Pylon, Granola, Posthog, Hubspot, Slack, and Notion. Poth connects that information into a living model of your customers, mapping analytics from Posthog and transcripts from Granola to the same customer, so you can finally answer the questions that live across multiple tools: churn, adoption or what to build next, with cited evidence. pothlabs.com**\n\n\n\n**The problem:**\n\nCompanies have customer information scattered across support tickets, sales calls, CRM notes, surveys, interviews, product analytics and more. \n\nBut answering a new question about customers still usually means manually searching across those systems or beginning another research project.\n\n**The (soft) pivot:**\n\nWe originally built Poth to generate hypotheses and launch adaptive interviews and surveys to test them. To determine what questions to ask and which customers to ask, Poth first had to understand all of the customer information a company already had.\n\nUnder the hood, we built a customer knowledge graph: Poth identifies the important objects in a company’s data: customers, accounts, users, products, features, conversations, behaviors, and feedback, and connects them through their relationships.\n\nAs we integrated with more teams, we noticed that several were using this underlying model as much as, or more than, the survey functionality itself. They wanted to ask questions across everything their customers had already said and done so we made that a core part of the product!\n\n**Customer knowledge graph example:**\n\nFor a company like Slack, Poth might model five objects:\n\n* **Account**  (a company paying for Slack)\n* **User**\n* **Workspace**\n* **Feature** (Slack Connect, Huddles, Canvas, or workflows)\n* **Customer signal** (A support ticket, call transcript, survey response, or product event)\n\nIt then connects them through relationships such as:\n\nUser —BELONGS_TO→ Account Account —OWNS→ Workspace User —USED→ Feature Customer Signal —CREATED_BY→ User Customer Signal —REFERENCES→ Feature\n\nSlack could then ask:\n\n“Why are enterprise accounts trying Slack Connect but not adopting it broadly?”\n\nAnd Poth could connect low Slack Connect usage in PostHog, permission-related support tickets in Pylon, security concerns mentioned in Granola transcripts, and renewal context from HubSpot, all through the same accounts, users, workspaces, and features.\n\nInstead of returning a generic summary, Poth can show the supporting and contradicting evidence, cite the underlying sources, and explain what information is still missing.\n\nIn practice, the model contains many more objects and relationships based on each company’s business and data.\n\n**Ask Poth:**\n\nConnect your customer data sources and internal documentation. Poth automatically builds the living model, and your team can ask questions naturally across everything customers have said and done.\n\nWhen the existing evidence is not enough, you can also use Poth to launch adaptive interviews and surveys to collect what is missing.\n\nSome use cases we’ve seen:\n\n* **Discovery:** “Across our discovery calls, what do users want most that we haven’t built?”\n* **Adoption and growth:** What is preventing activated customers from adopting more of the product and expanding?”\n* **Churn and risk:** “Identify the top three reasons customers have churned, then flag the five active accounts showing similar warning signs.”\n* **Win-loss analysis:** “What are the most common reasons we lose enterprise deals, and which product gaps are actually affecting revenue?”\n* **Account intelligence:** “Give me everything relevant to this account before my next customer call.”\n\n**Our ask:**\n\nWe’re already working with several teams in the YC batch, and we’re looking for more companies whose customer context is spread across several systems.\n\nI’d especially appreciate introductions to:\n\n* Founders and product leaders at growing startups with significant customer data across three or more tools\n* Heads of Customer Insights, Voice of the Customer, Customer Experience, User Research, or Customer Success at larger companies\n\nAnd if your team has important customer questions it still cannot answer confidently, email me at [matthew@pothlabs.com](mailto:matthew@pothlabs.com)! We’ll work closely with early teams to connect their core systems and get them live.\n\npothlabs.com", "url": "https://www.ycombinator.com/launches/RpS-ask-poth-ask-anything-about-your-customers", "votes": 7, "created_at": "2026-07-22T20:02:23.718Z" }, { "id": 103949, "title": "Poth Labs - The AI Native Customer Research Platform", "tagline": "Understand why users behave the way they do—hypotheses from your data, validated through interviews.", "body": "**TL;DR:** Poth helps you understand _why_ your users behave the way they do. It uses your company's data and context to generate hypotheses, then validates them through adaptive user interviews until it uncovers the root cause.\n\n**What we do:** Poth is an AI-native customer research platform\n\nAnalytics tools show behavior. Traditional research tools ask users static, predetermined questions. Poth connects the two—starting from your own data.\n\nIt takes a company's existing data and context to generate hypotheses, validates them through adaptive interviews, and continuously updates its questioning based on each user's responses—helping teams uncover the reasons behind churn, failed activation, checkout abandonment, or any user behavior.\n\n\n\n**Who we are:** Before Poth, I (Matthew) was an FDE at Palantir, where I worked closely with enterprise customers to help them make better use of their data. My co-founder Mojmir was a data scientist at Tietoevry, where he did the same.\n\nNow we're building Poth to help every company understand why users behave the way they do.\n\n**Our ask:** If you work in product, growth, customer research, or analytics, we'd love to chat — reach out and connect. Try it at [pothlabs.com](http://pothlabs.com).", "url": "https://www.ycombinator.com/launches/R2b-poth-labs-the-ai-native-customer-research-platform", "votes": 41, "created_at": "2026-06-25T19:41:41.476Z" } ] }, { "id": 31277, "name": "Dipole Labs", "slug": "dipole-labs", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/b476fcfad147bfb48b55b119e09431103b41f4d8.png", "website": "https://dipolelabs.com/", "all_locations": "Boston, MA, USA", "long_description": "", "one_liner": "AI-controlled optical switching for AI clusters", "team_size": 0, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1786560399, "tags": [ "Hard Tech", "Semiconductors", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/dipole-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/dipole-labs.json", "year_founded": 2026, "city": "Boston", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/b476fcfad147bfb48b55b119e09431103b41f4d8.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3178902, "full_name": "Deepankur Thureja", "title": "Founder/CEO", "founder_bio": "CEO & Co-founder @ Dipole Labs\n\nBuilding optical switching and compute modules that make AI clusters faster and more energy efficient.\n\nPreviously I was a postdoctoral researcher at Harvard where I developed advanced photonic and quantum devices, working at the intersection of materials science, optics, and AI.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dthureja", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ec3f292a37bbde4475aa4c89de8203153d747800.jpg" }, { "user_id": 3179632, "full_name": "Gabriele Pasquale", "title": "Founder/CTO", "founder_bio": "Co-Founder & CTO @ Dipole Labs\n\nBuilding optical switching and compute modules that make AI clusters faster and more energy-efficient.\n\nPreviously, I was a postdoctoral researcher at Harvard, where I developed advanced photonic and quantum devices, working at the intersection of materials science, optics, and AI.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/gabrielepasquale", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4ab80d1378669d55ada2deb72073e9390f4cbbae.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110374, "title": "Dipole Labs: Shooting lasers to solve the billion-dollar bottleneck inside AI datacenters", "tagline": "High-speed optical circuit switches that keep data moving as light, giving AI clusters up to 25% more useful compute from the same hardware.", "body": "**The most expensive computers ever built spend about half their time waiting for data.** As AI clusters scale, it's increasingly the network that limits how much compute you can actually use.\n\n**What we're building** Dipole Labs builds high-speed optical circuit switches for AI data centers. Between racks, data already travels as light, but routing it usually means converting it to electricity and back, burning power, generating heat, and adding latency. Dipole's switch keeps data in the optical domain, targeting sub-microsecond reconfiguration at large port counts, so the network can reshape itself around each workload.\n\nAn AI-native control layer learns GPU communication patterns and synchronizes optical reconfiguration with training and inference, turning the network into an active accelerator instead of a bottleneck.\n\n**Why it matters** The result: a 10,000-GPU cluster could perform like one with 12,500 GPUs, up to 25% more useful compute from the same hardware. At the scale of a 1 GW data center, that's up to $1 billion in additional compute capacity every year. Optical switching is just the start, the same platform extends from moving data as light to computing with it.\n\n**Who we are** Deepankur and Gabriele met as postdocs at Harvard and have spent the last decade building the photonic expertise behind this technology. Deepankur (CEO) built scalable quantum-optical hardware at ETH Zurich, published in Nature, and was awarded the ETH Medal for his doctoral thesis. Gabriele (CTO) has built nanophotonic and quantum devices at EPFL, and solved a decade-old condensed-matter physics problem published in Nature Materials, which earned the IBM prize from the Swiss and Austrian Physical Society.\n\n**Our ask** We're opening design partnerships with hyperscalers and neoclouds that want more useful compute from the hardware they already run, plus collaborations with photonic foundries to take our custom devices from prototype to scale. If that's you, let's talk.", "url": "https://www.ycombinator.com/launches/SiE-dipole-labs-shooting-lasers-to-solve-the-billion-dollar-bottleneck-inside-ai-datacenters", "votes": 9, "created_at": "2026-08-13T17:30:00.146Z" } ] }, { "id": 31298, "name": "Applied Electrodynamics, Inc.", "slug": "applied-electrodynamics-inc", "former_names": [ "Applied Electrodynamics" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/11648bf43e6f51e66e8e798d40073877c53c2f20.png", "website": "https://www.ae-dyn.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Applied Electrodynamics builds imaging systems that see through solid matter. Our team previously co-invented the iPhone touchscreen and Nest learning thermostat. We’re starting in construction, helping teams map structural elements and utilities inside walls and ceilings. The same platform extends to security screening and cargo inspection. Our mission is to make the hidden physical world machine-readable across the entire electromagnetic spectrum.\r\n", "one_liner": "A new kind of camera that can see through walls.", "team_size": 5, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1785202775, "tags": [ "Hard Tech", "Hardware", "Construction", "Manufacturing", "Defense" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/applied-electrodynamics-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/applied-electrodynamics-inc.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/applied-electrodynamics", "x_url": "https://x.com/ae_dynamics", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/11648bf43e6f51e66e8e798d40073877c53c2f20.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 2724984, "full_name": "Bilgehan Avser", "title": "Founder/CEO", "founder_bio": "Bil Avser previously led antenna technologies at Apple across multiple iPhone generations, including UWB, 5G, and mmWave systems, before serving as Head of Technology at Humane through its acquisition by HP. He holds a PhD in Electrical Engineering, has 35+ patents in electromagnetics and RF systems, and received the IEEE APS Pre-Doctoral Research Award.\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/bavser", "x_url": "https://x.com/bavser", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1b1a0995ed1bd232a77653db4e4dee2ab647ac71.jpg" }, { "user_id": 2980661, "full_name": "Brian Huppi", "title": "Founder", "founder_bio": "Brian Huppi is a hardware engineer and inventor who helped build the earliest iPhone prototypes at Apple, contributing to the multitouch hardware that redefined how the world interacts with devices. He co-invented the first generation Nest Learning Thermostat and later worked on the AI Pin at Humane. He is a named inventor on numerous patents spanning touch interfaces, sensing systems, lighting, and connected devices. Brian holds degrees in Civil and Electrical Engineering from UMass Amherst.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/brian-huppi-69a8433/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d67b0681b838d1fed82f35b18d6a43b7bebae7ab.jpg" }, { "user_id": 2974809, "full_name": "Paul Leutheuser", "title": "Founder", "founder_bio": "Paul Leutheuser is a hardware product design engineer who drove numerous new iPhone features into production, including the first haptic home button and IPX8 sealed enclosure. He then designed and shipped multiple 0-1 products at Humane, leading the Product Design team post-acquisition. Paul has 11 patents and holds a Mechanical Engineering degree from UC Berkeley.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/paul-leutheuser-54207364/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/034c1ff34b8c99a04807aa712b02f94f3c6d3e72.jpg" }, { "user_id": 2980566, "full_name": "George McLean", "title": "Founder", "founder_bio": "George McLean is an applied physicist with wide-ranging experience in semiconductor imaging systems and medical devices and patents in retinal prothesis and wearable physiological monitoring. At Humane he made key contributions to the laser display of the Ai Pin. He holds degrees in physics from MIT and UC Berkeley.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/george-mclean-75990415/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4b8c4f90c563da550a00fdb0e86935250883a24f.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108275, "title": "Applied Electrodynamics: A new kind of camera that can see through walls", "tagline": "Making the physical world machine-readable across the entire electromagnetic spectrum", "body": "**TL;DR:** We are building a new kind of camera that can see through walls using radio waves. We provide construction, industrial inspection, and security teams with an accurate map of hidden objects behind walls and composite surfaces.\n\n\n\n## **The Problem**\n\nMost of the built world is opaque.\n\nPipes, cables, and structural elements hide behind walls until they're discovered mid-project, driving $20B+ a year in avoidable rework.\n\nDefects remain hidden inside composites until parts break.\n\nEntry teams are missing critical information before they breach a barrier.\n\n## **The Solution**\n\nAn imaging platform that uses radio waves to see behind walls. It works just like a camera, requiring no training and producing high-resolution images that are instantly understandable. The resulting point cloud is natively to scale, accurate down to 1mm.\n\n## **The Team**\n\nFour engineers who've worked together for 8+ years, shipped multiple zero-to-one hardware products, and hold 250+ patents.\n\n**Bilgehan** — PhD in electromagnetics; led antenna technologies on four iPhone generations, shipping the first phones with UWB and 5G mmWave; inventor of the zirconia dielectric resonator antenna.\n\n**Brian** — built the earliest iPhone prototypes at Apple, co-invented iPhone multitouch and the Nest Learning Thermostat; degrees in Electrical and Civil Engineering.\n\n**George** — MIT physicist; built x-ray imagers, retinal implants, optical sensors and the world's smallest laser projector.\n\n**Paul** — Berkeley Mech.E.; ex-Apple and Humane product design lead; co-invented iPhone haptic Touch ID, IPX8 sealing.\n\n## **What we built**\n\n![uploaded image](/media/?type=post&id=108275&key=user_uploads/2724984/c00b3438-c7f7-433f-941d-e021791e51c5)\n\nWe've built the hardware and imaging algorithms that can see behind composite layers and common building materials (sheetrock, tiles, marble, quartz, brick and insulation). Next step is our handheld platform that fuses our radio imaging with RGB and thermal cameras.\n\n![uploaded image](/media/?type=post&id=108275&key=user_uploads/2724984/f1cf1ffa-8b23-4780-b363-13b0db90832d)\n\n## **Our Ask**\n\nWe're looking for intros to decision makers at general contracting firms and industrial inspection teams, and to contacts with military access who can introduce and demo our systems to US defense organizations.\n\nIf that's you or someone you know, reach out to us at [founders@ae-dyn.com](mailto:founders@ae-dyn.com)", "url": "https://www.ycombinator.com/launches/SAN-applied-electrodynamics-a-new-kind-of-camera-that-can-see-through-walls", "votes": 25, "created_at": "2026-07-30T14:48:05.478Z" } ] }, { "id": 31301, "name": "OpenVector", "slug": "openvector", "former_names": [ "TeleOps", "BotZero", "Foveal" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/54df715903ce304e2b9e3b2fe1f375caf7811ea6.png", "website": "https://openvector.com", "all_locations": "San Francisco, CA, USA", "long_description": "OpenVector transforms any existing camera into an automated 24/7 workforce without requiring new hardware. We've built the world's fastest vision-language model inference engine that allows users to simply describe a task in plain English and automatically build a custom AI vision model to monitor their feed. When it detects an important physical event—like a stalled production cycle, a safety violation, or a waiting customer—it operates a computer just like a human to take immediate action. It integrates with legacy systems, makes calls, sends emails, updates a POS/CRM, or triggers webhooks in real time. OpenVector enables any camera to do real work.", "one_liner": "AI that turns any camera to an autonomous worker with real-time vision", "team_size": 4, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1785369468, "tags": [ "Robotics", "Computer Vision", "Infrastructure", "Industrial" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/openvector", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/openvector.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://linked.com/in/openvectorlabs", "x_url": "https://x.com/openvectorlabs", "fb_url": "https://www.facebook.com/profile.php?id=61591381283372", "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/54df715903ce304e2b9e3b2fe1f375caf7811ea6.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 633157, "full_name": "Andrey Gizdov", "title": "Founder/CEO", "founder_bio": "Dropped out of PhD. Computer Vision @ Harvard.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/andreygizdov/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/994295941ffb048fecb964921a35db1494d051b3.jpg" }, { "user_id": 2384953, "full_name": "Vishal Urlam", "title": "Founder", "founder_bio": "CMU'23, ECE'20 | MIT GrandHack '25 | Kellogg Design Challenge '23", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/vishal-urlam/", "x_url": "https://x.com/vishalurlam", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/398802b070ebc9a50a9873e3441541de12ccedfb.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108360, "title": "OpenVector - Self-improving AI that turns any camera into an autonomous worker.", "tagline": "Most cameras just record. We give them the ability to do real work.", "body": "**TLDR;** OpenVector turns existing cameras into an automated 24/7 workforce. Describe a task in plain English, and it auto-builds a custom vision model that takes real-time actions on a computer when it spots an event.\n\n**Ask;** If you or someone you know runs physical operations (manufacturing, logistics, retail, security) and wants to automate visual tasks without buying new hardware, we’d love to connect!\n\n**Our Launch Video:**\n\n\n\n**Our Team:**\n\nOpenVector was founded by a team of builders and engineers obsessed with bringing self-improving visual intelligence to the physical world.\n\n* [Andrey Gizdov](https://www.linkedin.com/in/andreygizdov/), CEO, dropped a fully-funded **PhD** scholarship from Fulbright, worth **$170,000**, with invitations at **Harvard and MIT**, where he worked on real-time vision models.\n* [Vishal Urlam](https://www.linkedin.com/in/vishal-urlam/), CTO, previously automated camera systems and network protocols for Tata, the **second-largest network infrastructure in the world**. CMU embedded systems engineer.\n\n**Problem**:\n\nMany important business moments happen in the physical world: Someone tailgates. A production cycle stalls. A worker skips a step in the manufacturing process. A customer arrives, and no one responds. There are over 1 billion cameras worldwide, but they only record it for later. For physical operations, the best technology available to catch these events in real-time has been a human staring at a monitor.\n\n**Solution:**\n\nOpenVector is an AI that brings action to any camera.\n\n* **Plain English to AI Model:** No coding or expensive setup. Simply describe your task in plain English, and OpenVector automatically builds the best AI vision model for it, and watches your camera feeds 24/7. It’s an app that takes 2 minutes to download and set up, and works with any camera.\n* **Acts Like a Human:** OpenVector operates a computer just like a human based on what it sees.\n* **Legacy Integration:** It automatically integrates with your existing systems to make a call, send an email, update your POS or CRM, or trigger a webhook the second an event happens.\n* **Zero New Hardware:** Works entirely with the cameras you already have installed (custom end-to-end solutions also available).\n\n**We’ve Already Started:**\n\nOpenVector is already live and seeing use cases across security, retail, manufacturing, and logistics. We are tracking production cycle efficiency directly from existing CCTV, ensuring good-manufacturing-practices, and catching physical design errors.\n\n**Now Onboarding:**\n\nOpenVector is ready to deploy on your camera feeds today. Contact us if you’re tired of cameras that only watch and want them to actually do work for you. [founders@openvector.com](mailto:founders@openvector.com)\n\n![uploaded image](/media/?type=post&id=108360&key=user_uploads/633157/46faabca-90e0-4c69-a364-138588d733d5)\n\n", "url": "https://www.ycombinator.com/launches/SBk-openvector-self-improving-ai-that-turns-any-camera-into-an-autonomous-worker", "votes": 16, "created_at": "2026-07-31T18:00:43.590Z" } ] }, { "id": 31304, "name": "Litmus", "slug": "litmus-hiring", "former_names": [ "Avert", "Litmus", "Avert" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/4a9f14022a002a449869ff376c950e1a9966d580.png", "website": "https://litmushiring.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Litmus is building the most accurate framework for evaluating and benchmarking human capability, starting with software.\r\n\r\nAI will compound small differences in human capability into increasingly large differences in what people can accomplish, while making existing static benchmarks obsolete. Software is already there: AI can hill-climb any output-based evaluation, while the ability to direct it is becoming the defining advantage. \r\n\r\nLitmus applies the same approach we already use for model evals to humans – creating world-like environments, and inspecting trajectory instead of just output.\r\n\r\nEvery knowledge industry will soon face the same problem. We build Litmus to tell you what humans are capable of.\r\n\r\nWe're already helping build frontier technical teams at Mercor, Composio, Neo Scholars, and more.", "one_liner": "Evals for humans", "team_size": 3, "industry": "B2B", "subindustry": "B2B", "launched_at": 1783537104, "tags": [ "Recruiting", "Data Engineering", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/litmus-hiring", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/litmus-hiring.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/litmushiring", "x_url": "https://x.com/UseLitmus", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/4a9f14022a002a449869ff376c950e1a9966d580.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/7d94ea3ae5f80a2b9c035af7e2b630fc5b19556e.png" ], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1353649, "full_name": "Shaivi Rau", "title": "Founder/CEO", "founder_bio": "building litmus", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shaivi-rau", "x_url": "https://x.com/ShaiviRau", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d5482bbda470c38ccb104e539a75d967d59855d2.jpg" }, { "user_id": 2315600, "full_name": "Elena Zhao", "title": "Founder/CTO", "founder_bio": "Evals for humans", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/elena-zhao-015353217/", "x_url": "https://x.com/ElenaxZhao", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/181a9fd30489864e7baeee3f884bda9fe2c72cf1.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 105006, "title": "Litmus - The industry's most accurate engineering interview", "tagline": "Scalable, async work trials built from your repos, tickets, and JDs", "body": "**TL;DR:** We evaluate candidates through async technical assessments built from your own context (repos, tickets, and job descriptions) and let them work in their own setup with all their tools while we track and evaluate their process and output. _Engineering is changing; we help you hire for the role it’s becoming. Learn more at [litmushiring.com](http://litmushiring.com)._\n\n\n\n**The problem**\\\nYou're a month in with a new hire and they're just not getting it: struggling to ship independently, shipping AI-generated code they can't explain, clueless when production breaks. None of this showed up in your interview because your interview was never designed to catch it.\n\nWe all already agree LeetCode sucks. But work trials are time-draining, unscalable, and kind of awkward when you know within the first hour someone's not gonna make it. So most teams are stuck choosing between a broken screen and a process that doesn't scale.\n\n**What we built**\\\nLitmus fixes both. We turn your repos, tickets, and job descriptions into a full interview pipeline where every assessment is generated from scratch for your company. Not a generic template, your actual codebase, your real engineering problems, scoped as end-to-end features.\n\nCandidates work in their own setup, with all their tools including AI because that's how engineering works now, and that's what you're hiring for. \n\nEvery submission runs in a sandbox, so graders evaluate real code execution, not what looks good on paper, and grading is tuned to your company across tool fluency, process, and outcome. \n\nThe result: a real read on how someone actually builds, not just how they interview.\n\n1. **Generate Tailored Assessments**\n\n ![uploaded image](/media/?type=post&id=105006&key=user_uploads/1353649/15424c0c-1ed2-4c6d-ac82-98ec2d81f8f1)\n\n2. **Candidates take Assessments**\n\n ![uploaded image](/media/?type=post&id=105006&key=user_uploads/1353649/af3edf38-52ab-4a83-a5c5-caeb40c8a532)\n\n3. **Evaluate at Scale**\n\n ![uploaded image](/media/?type=post&id=105006&key=user_uploads/1353649/8b8a71a8-44c3-408d-ba04-80cde5f997bf)\n\n4. **Deep Dive into Top Candidates**\n\n ![uploaded image](/media/?type=post&id=105006&key=user_uploads/1353649/696687d6-40b7-4c13-ac4b-05d60d6f54df)\n\n\n**Traction**\\\n**3,000+ assessments evaluated**. **$60k ARR**, growing fast. We've built the hiring system for **Neo Scholars**, their portfolio, and others. Let us bring our expertise to your team.\n\n**Who we are**\\\nWe're Shaivi and Elena. As a SWE at Two Sigma and Meta, Elena watched people ace every interview and then fail to build anything real on the job. Shaivi worked at early-stage startups and in VC, sitting beside founders as they agonized over which engineers to trust with the product. Two sides of the same coin - the most important decision a company makes is its team, and current interviews still can't tell you who'll deliver. Litmus does.\n\n**Our ask**\\\n**Scaling your eng team?** **We'll come to your office, set you up, and run Litmus on your next role. Book a demo at [litmushiring.com](http://litmushiring.com).** And if you know a team drowning in engineering interviews, send them our way.", "url": "https://www.ycombinator.com/launches/RJe-litmus-the-industry-s-most-accurate-engineering-interview", "votes": 5, "created_at": "2026-07-09T15:01:22.190Z" } ] }, { "id": 31340, "name": "Illume Labs", "slug": "illume-labs", "former_names": [ "Illume" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/72dfae7e6f3aa505626a7c902f014e294cc6ab91.png", "website": "https://www.illumelabs.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Illume is a 24/7 AI personal health & longevity companion that connects your wearables, blood panels, and genomic data to text you personalized and actionable insights. We pull in your health data and fuse it into one secure model. You get daily text updates on what changed, what it means, and what to do next; conversational health Q&A like \"What should I do to improve my bloodwork?\"; cross-data trend detection that catches patterns individual platforms miss; and a full platform to dig deeper, explore your health data, and text your Illume agent anytime. We're 3 MIT grads & friends, with experience at Windsurf, MIT CSAIL, Eli Lilly, & the Broad Institute. ", "one_liner": "24/7 Personal Health & Longevity Companion", "team_size": 3, "industry": "Healthcare", "subindustry": "Healthcare -> Consumer Health and Wellness", "launched_at": 1784104321, "tags": [ "Generative AI", "Health Tech", "Biotech", "Health & Wellness", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare", "Consumer Health and Wellness" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/illume-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/illume-labs.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/111286108/", "x_url": "https://x.com/IllumeLabsInc", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/72dfae7e6f3aa505626a7c902f014e294cc6ab91.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/b71f245c30e326bd5ab760878796f54eb60e13a9.png" ], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 2698620, "full_name": "Pari Latawa", "title": "Founder", "founder_bio": "Founder at Illume Labs (S26)\nMIT BS + MEng in CS & Biology", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/paridhilatawa/", "x_url": "https://x.com/parilatawa", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b322dddf9ddbaf085c99c79b6237dca80adf97cb.jpg" }, { "user_id": 1603152, "full_name": "Frank Lee", "title": "Founder", "founder_bio": "Founder at Illume Labs (S26). CS + Math @ MIT.\nPreviously software engineer at Windsurf and researcher in Pulkit Agrawal Lab @ MIT.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/frank-lee-/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8b42be2a9d604ce6408a43515cd8cc10ad119308.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 106534, "title": "Illume Labs - Your 24/7 personalized health companion", "tagline": "Illume is a 24/7 personal health companion that connects your wearables, nutrition, workouts, and lab results to give you personalized, actionable insights over text. ", "body": "**Hey everyone 👋!** We’re [Pari](https://www.linkedin.com/in/paridhilatawa/), [Frank](https://www.linkedin.com/in/frank-lee-/), and [Jin](https://www.linkedin.com/in/jinwong01/), co-founders of Illume. \n\n**TL;DR:** Illume is a 24/7 personal health companion that connects your wearables, nutrition, workouts, and lab results to give you personalized, actionable insights over text. \n\n**Join our private beta waitlist: [www.illumelabs.ai](http://www.illumelabs.ai). \\\nWatch Our [Launch Video](https://www.youtube.com/watch?v=bsfFOzB0Nho):**\n\n**Problem:** \n\nHealth data is more abundant than ever, but it remains scattered. Nearly six in ten U.S. adults own a wearable or connected health device, and nearly one in three used an AI chatbot for health questions. People are generating more health data than ever and actively looking for help understanding it, but their data and guidance still live in separate places. \n\nWe saw this firsthand as family members struggled to piece together glucose readings for diabetes and blood tests for cardiovascular risk, all while juggling a pile of fitness, nutrition, and health-tracking apps that never talked to each other. \n\nPeople increasingly turn to ChatGPT or Claude for answers, but without continuous context to their health, the advice is often generic, incomplete, and difficult to track over time.\n\n**Our solution:** \n\nIllume creates one home for your health and gives you a personal health companion you can text. \n\nToday, you can: \n\n* Connect your wearable and automatically sync sleep, activity, and recovery data \n* Text Illume photos of your meals to log and understand your nutrition \n* Record workouts and track your training \n* Upload bloodwork and lab results \n* Ask questions about your health using the context of your own data \n* Receive proactive insights about what changed, what it might mean, and what to do next\n\nFor example, you can ask: \n\n* “Why am I so exhausted today? Make me a dashboard of all the relevant metrics!” \n* “How has eating late affected my sleep?”\n* “Am I recovering well enough to train? Run a regression!”\n* “What should I focus on before my next blood test?”\n\nIllume connects patterns across your data that individual platforms can’t see, giving you a more complete picture of your health instead of another isolated score or dashboard. \n\nYou can text your Illume agent for quick answers or use the platform to explore your data, trends, and history in greater depth.\n\n**Our mission**\n\nOur long-term vision is to make healthcare more proactive, continuous, and personalized. \n\nBy combining continuous biometrics with nutrition, training, blood panels, and eventually genetic risk, Illume could help surface meaningful changes earlier and help people understand their susceptibility to conditions such as cardiovascular disease, diabetes, and Alzheimer’s. \n\nWe want to help people move from asking, “What happened?” to understanding what may be changing and what they can do about it. \n\n**Who’s it for:** Illume is initially designed for people who are proactive about their health and longevity- especially lifters, biohackers, and anyone already using wearables, blood panels, CGMs, nutrition trackers, or genetic testing.  \n\n**Who we are:** We’re three MIT graduates who met in freshman year! \n\n[Pari Latawa](https://www.linkedin.com/in/paridhilatawa/): MIT BS & MEng in CS + Molecular Biology. Previously conducted research at the Broad Institute, Eli Lilly, & MD Anderson. Founded EasyComm to build tools for children with Down Syndrome, and SparkED, a health-literacy nonprofit.\n\n[Frank Lee](https://www.linkedin.com/in/frank-lee-/): MIT BS in Computer Science and Mathematics. Conducted RLAIF research for language models at MIT CSAIL and previously built agent systems at Windsurf.\n\n[Jin Wong](https://www.linkedin.com/in/jinwong01/): MIT BS in Business Analytics. Previously conducted research at the MIT Media Lab and deployed coding agents at Windsurf. \n\n**Our ask:** \n\n**Join the private beta waitlist at [www.illumelabs.ai](http://www.illumelabs.ai).** We’ll invite an initial group of founding members to receive early access and help shape the product. \n\n**If you want to chat or can intro us** to someone in training, sports, longevity, and/or biohacking, email us at [founders@illumelabs.ai](mailto:founders@illumelabs.ai). We’d love to stop by, chat, and show you how Illume works! ", "url": "https://www.ycombinator.com/launches/RiI-illume-labs-your-24-7-personalized-health-companion", "votes": 28, "created_at": "2026-07-20T17:02:49.112Z" } ] }, { "id": 31341, "name": "Most Robotic", "slug": "most-robotic", "former_names": [ "Instance" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/16ccadcbd2ec474ed470efbc872763f4a7984ec0.png", "website": "https://mostrobotic.com/", "all_locations": "San Francisco, CA, USA", "long_description": "We're a robotics research lab pushing what's possible with physical AI.\r\n\r\nWe're MIT computer scientists & friends since middle school, with robotics research backgrounds and industry experience at NASA JPL, SpaceX, and AWS.", "one_liner": "Research lab building the most robotic household in the world", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1782761102, "tags": [ "Robotics", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/most-robotic", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/most-robotic.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/mostrobotic/", "x_url": "https://twitter.com/mostrobotic", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/16ccadcbd2ec474ed470efbc872763f4a7984ec0.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/5d0a5df4551ace828d9502da4d71b644d4143a15.png" ], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 3107325, "full_name": "Claire Mao", "title": "Founder/CEO", "founder_bio": "MIT Math + CS. Aerospace research @ MIT Media Lab; prev @ NASA JPL", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/clairemao-/", "x_url": "https://x.com/clairemao78", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8bc8742026b7ae3d795367ca9cbc052f0035d9c3.jpg" }, { "user_id": 2386225, "full_name": "Lucy Cai", "title": "Founder/CTO", "founder_bio": "MIT CS + AI. Robotics research @ MIT LIS Lab; prev @ SpaceX, Amazon (AWS)", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/lucy-cai/", "x_url": "https://x.com/LucyCai9", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/7e3c58d6a8b486439a545ea902cf0dd6981d51ed.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 105382, "title": "Instance: Automated evaluation for robot policies, starting with the success detector", "tagline": "Today, evaluating a robot policy means humans have to watch the robot roll out, mark success, and reset the scene – we're automating that.", "body": "**TL;DR:** We’re **automating robot evaluation**– the loop where you wait for your robot to roll out, judge its success, and then reset the scene. Today, we’re launching step one: a **success detector** that automatically tells you whether your robot actually completed its task. Book a demo with [founders@instancelabs.ai](mailto:founders@instancelabs.ai) and we’ll visit your office to show you how it works!\n\n\n\n**The problem**\n\nIn Lucy’s robotics research this past year, she **spent way too much time being a babysitter for her robot**. Her project was to get a humanoid to find objects around a room, so her days went like this: start a rollout, watch the robot open all the drawers in the room for 30 minutes, mark success/fail in a spreadsheet, then get up, close all the drawers, and put the objects back where they started. Then do it again, and again, and again. Some days ended with back pain, and for weeks she couldn’t work on anything else because she was stuck doing exactly this.\n\nThat was one robot, in one lab, running a single policy. Industry teams run fleets of robots across dozens of policies, with **multiple people assigned to sit and watch each one**. Every robotics team we’ve talked to has some version of this– and it will only grow as robot foundation models scale.\n\n**Our solution**\n\nInstance is building the **autonomous evaluation rig for robot learning**: your robot rolls out, we judge success, a second robot resets the scene, and the next rollout starts– with no humans in the loop at all.\n\nToday, we’re starting with the success judge. Point us at a task description and your video rollouts, and we return a verdict with detailed subtask captions. We benchmarked our verifier against 10,000+ held-out, human-labeled episodes across 7 robot platforms, and it resulted in **higher accuracy than Claude Opus 4.8, at a fraction of the latency.**\n\n**Try it**\n\nTry it at [demo.instancelabs.ai](http://demo.instancelabs.ai) – drop in a video, or record one from your phone and watch it get verified live.\n\n**Team**\n\nWe’re Claire and Lucy, MIT grads and friends since middle school! \n\nClaire studied math and CS at MIT. At NASA JPL, she built software to simulate planetary atmospheres, and at the MIT Media Lab, she researched sustainable propulsion for spaceflight. She loves working where software meets the physical world.\n\nLucy studied CS at MIT and completed her MEng in MIT's Learning and Intelligent Systems lab. She’s an experienced robotics researcher, and has in the past worked on satellite software at SpaceX, LLM unit testing at Amazon, and brain computer interfaces at Blackrock Neurotech.\n\n**Our ask**\n\nIf you are:\n\n* training robot policies\n* running evals\n* deploying robots in the real world\n* or can intro us to someone who is,\n\nEmail us ([founders@instancelabs.ai](mailto:founders@instancelabs.ai)) to book a demo. We’d love to swing by your office with sweet treats to show you how it works!\n\n![uploaded image](/media/?type=post&id=105382&key=user_uploads/3107325/d03f1074-ae6e-4ab3-acdd-81ff69be93b6)\n\n", "url": "https://www.ycombinator.com/launches/RPi-instance-automated-evaluation-for-robot-policies-starting-with-the-success-detector", "votes": 36, "created_at": "2026-07-13T20:48:30.605Z" } ] }, { "id": 31349, "name": "Palette", "slug": "palette-2", "former_names": [ "Iris Prism" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c4b357efedd6b75b24594234b97cea3bade805e0.png", "website": "https://palettelabs.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Palette is a self-improving AI-native media platform built by MIT computer scientists and multimodal researchers that unifies the creative workflow into a single generation engine. We’re building a fully autonomous production engine that generates, distributes, and optimizes interactive media for every brand and audience.", "one_liner": "The Media Engine of the AI Era", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1783722381, "tags": [ "Generative AI", "Video", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/palette-2", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/palette-2.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/palette-labs-1", "x_url": "https://x.com/trypalette_ai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c4b357efedd6b75b24594234b97cea3bade805e0.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 1575309, "full_name": "Josephine Lee", "title": "Founder", "founder_bio": "CEO @ Palette (YC S26) | CS @ MIT", "is_active": true, "linkedin_url": "https://linkedin.com/in/josephinelee1234", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/7a8529ef3e65940ccbd144092bcd1702bb8a23de.jpg" }, { "user_id": 2687428, "full_name": "Serena Pei", "title": "Founder", "founder_bio": "CTO @ Palette (YC S26) | CS @ MIT", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/serena-pei-80ab21215/", "x_url": "https://x.com/spei04", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fff58978928940e0fd7ca7ed61c7181fd027c4a4.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108469, "title": "Palette - An AI-Native Media Platform", "tagline": "Unify the creative workflow into a single, high quality generation engine", "body": "[Palette](https://palettelabs.com/) is an AI-native media platform built by MIT computer scientists and multimodal researchers that unifies the creative workflow into a single generation engine. We eliminate tool and context fragmentation for creative teams in enterprises and allow content creators to transform static assets into controllable, repeatable workflows in one conversation.\n\nCheck out our demo [here](https://youtu.be/xKibxXF9pGw)\n\n**The Problem**\n\nToday, in order to deliver high-quality, tasteful content, individual content creators and enterprise teams struggle with a heavily fragmented landscape of separate, disconnected applications for multimodal content production. For individual creators, this tool fragmentation breaks creative flow and focus. For enterprise teams, it destroys brand context and causes inconsistencies that slow down iteration speed.\n\n**The Solution**\n\nPalette solves this by providing a single platform where content creators can use natural language and product context to generate multimodal media, execute precise edits, and build repeatable creative workflows. To drive enterprise growth, Palette delivers targeted, end-to-end services, ranging from ad runs and brand-safe influencer UGC to product launches and interactive multimedia L&D onboarding.\n\n**About us**\n\nWe’re [Josephine](https://www.linkedin.com/in/josephinelee1234/) and [Serena](https://www.linkedin.com/in/serena-pei-80ab21215/), the founders of Palette. We met at an art program and studied computer science at MIT, with experience in multimodal research and computer vision as well as industry experience at Scale AI, Datadog, and Amazon. \n\n**Asks**\n\nWe’d love to connect with:\n\n* Content creators looking to explore precise, controllable visual content generation.\n* Marketing teams looking to scale their content and advertisement workflows.\n* Operations leaders facing severe onboarding or documentation bottlenecks.\n* Heads of Learning & Development looking to move away from static LMS systems.\n\nTry it at [studio.palettelabs.com](http://studio.palettelabs.com) or learn more at [palettelabs.com](http://palettelabs.com).", "url": "https://www.ycombinator.com/launches/SDV-palette-an-ai-native-media-platform", "votes": 36, "created_at": "2026-07-31T19:02:01.948Z" } ] }, { "id": 31352, "name": "Marker", "slug": "marker", "former_names": [ "Widari", "Widari Inc." ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/93931f5e81e962ad8d0a22dae5ec988c5226fb07.png", "website": "https://onmarker.com", "all_locations": "San Francisco, CA, USA; Remote", "long_description": "Marker is a platform for rebuilding enterprises agent-first, delivered by forward-deployed engineers. It productizes every step from discovery to production, so every agent we ship makes the next one faster.", "one_liner": "Platform + FDEs that rebuild enterprises agent-first", "team_size": 3, "industry": "B2B", "subindustry": "B2B -> Productivity", "launched_at": 1787016991, "tags": [ "Generative AI", "SaaS", "B2B", "Workflow Automation", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Productivity" ], "regions": [ "United States of America", "America / Canada", "Remote", "Fully Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/marker", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/marker.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/112296703/", "x_url": "https://x.com/markerhq", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/93931f5e81e962ad8d0a22dae5ec988c5226fb07.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 334251, "full_name": "David Pastewka", "title": "Founder", "founder_bio": "Co-founder & CEO of Marker. Previously co-founded Drapr (S20), acquired by Gap Inc. UC Berkeley, Business & Computer Science.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/davidpastewka/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a6a89814f4f0727febfa36792bba3bf24dff3432.jpg" }, { "user_id": 35831, "full_name": "Richard Berwick", "title": "Founder", "founder_bio": "Co-founder of Marker. Previously co-founded Drapr (S20; acquired by Gap Inc.) and Twindom (acquired). UC Berkeley, Business.\n\nEntrepreneur at heart with a hacker brain. I'm always looking for ways to solve problems in efficient, inexpensive ways -- and I love helping people (and companies) get through difficult situations. If I can create value, I will.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/richard-berwick-93995629/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/feaf807d92184710fb01168ec1224a14a4583a1c.jpg" }, { "user_id": 35832, "full_name": "Will Drevno", "title": "Founder", "founder_bio": "Co-founder at Marker. Previously co-founded Drapr (yc s20; acquired by Gap Inc.) and Twindom (acquired). Studied engineering at UC Berkeley.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/will-drevno-0a78832a/", "x_url": "https://twitter.com/WillDrevno", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/49050f23e7b38cb215076ab6eaa87155b8076252.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111180, "title": "Marker: AI-native consultancy rebuilding businesses with AI agents", "tagline": "95% of enterprise AI initiatives fail. Ours don't.", "body": "**TL;DR: Marker partners with execs, learns how their organization actually works, finds the highest-impact problems, and solves them by** **shipping custom AI agents into production on the Marker platform in weeks. Each agent deepens our understanding of the business, so the next agent launches faster. We stay embedded until the business is rebuilt.**\n\n\n\nWe’re Will Drevno, Richard Berwick, and David Pastewka, the founders of Marker. We’ve built together for 15 years. Our last company, Drapr (YC S20), was acquired by Gap Inc., and Richard went on to found Gap’s Office of AI.\n\n## **Why we built Marker**\n\nInside Gap, we saw firsthand how much leaders wanted to transform their businesses with AI and how hard it was to turn that ambition into production systems that actually changed how their organizations worked.\n\n**95% of AI initiatives still fail. Ours don’t.**\n\nThe hard part is understanding how work happens across people, systems, and teams; choosing the opportunities that matter most; and getting solutions into production.\n\n## **How Marker works**\n\nWe embed with each company and lead the transformation end to end:\n\n* **Understand the organization.** We listen closely to leaders and frontline teams, learn their systems and ways of working, and map the processes that usually live in people’s heads.\n* **Find the biggest opportunities.** We identify and prioritize the problems where AI can have the greatest impact on cost, capacity, speed, or growth.\n* **Build and ship.** We create custom agents and automations around the company’s real workflows and systems. The agents we deploy run on the Marker Command Center and go live in production within weeks.\n* **Compound.** Every agent we deploy deepens our understanding of the company. That helps us launch the next one faster and makes Marker increasingly integral to how the organization operates.\n\n## **Who we’re working with today**\n\nWe’re rebuilding companies from mid-size through Fortune 500.\n\n## **Why us**\n\nWe’ve sat on every side of this: founders selling into enterprises, operators transforming one from the inside, and executives responsible for enterprise AI. Marker is the company we wish we could have hired at Gap.\n\n## **Our ask**\n\n**If you are (or you know) a leader looking for real AI transformation—not another pilot, deck, or self-serve tool—please send them our way.**\n\nReach us via email at [**founders@onmarker.com**](mailto:founders@onmarker.com), or visit [**onmarker.com**](http://onmarker.com).", "url": "https://www.ycombinator.com/launches/SvE-marker-ai-native-consultancy-rebuilding-businesses-with-ai-agents", "votes": 72, "created_at": "2026-08-19T14:00:01.019Z" } ] }, { "id": 31425, "name": "WonderTx", "slug": "wondertx", "former_names": [ "A Company in Stealth Mode Incorporated" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/08d90d66366f5066cf3ad5ae02f156dc4bcc8ebe.png", "website": "https://wondertx.ai/", "all_locations": "San Francisco, CA, USA; Remote", "long_description": "We’re an AI-native biotech, working on replacing needles with pills. Many drugs, like insulin or Ozempic, are wonderful except you have to inject them. We discover new molecules that you can take as a pill instead. Because the pills operate on biology which has been shown to be effective in people, rather than merely in animal models of disease, we avoid what has historically been the single biggest risk in drug discovery.  This should dramatically increase our probability of success.\r\n\r\nThere’s $200B of treatable diseases and validated human biology without a pill because discovery programs lack a tractable chemical starting point. This is a zero-shot inference problem, because these targets have no training data, so LLMs and other historical interpolative AI approaches are insufficient. Instead, we are building our own extrapolative frontier models that reason beyond their training data. Most critically, we have repeatedly validated our platform on both our own and partner programs. On 4 biologically- and structurally-diverse targets where we had zero training data, we successfully picked binders that were validated experimentally, with confirmation ranging from primary screen hits to crystallographic structures.", "one_liner": "Extrapolative AI to unlock First-in-Class drugs", "team_size": 3, "industry": "Healthcare", "subindustry": "Healthcare -> Therapeutics", "launched_at": 1786648812, "tags": [ "AI-powered Drug Discovery", "Generative AI", "Biotech" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare", "Therapeutics" ], "regions": [ "United States of America", "America / Canada", "Remote", "Fully Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": [ { "question": "What is your long-term vision? If you truly succeed, what will be different about the world?", "answer": "Imagine how drug discovery is done 80 years from now and it’s obvious that AI is going to be woven through every piece and every step.  No one believes that we will have less data, or that it will be easier to understand, or that we won’t need advanced statistical algorithms, or that the computers will be less capable. There is an inevitability here. But let’s not wait 80 years! If we succeed then the path from conception to clinic for the toughest diseases, for the most difficult biological problems, becomes tractable. Our practical outcome is simple: patients whose diseases cannot currently be drugged effectively or at all will live better longer lives. " } ], "url": "https://www.ycombinator.com/companies/wondertx", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/wondertx.json", "year_founded": null, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/wondertx/", "x_url": "https://x.com/WonderTxAi", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/08d90d66366f5066cf3ad5ae02f156dc4bcc8ebe.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Pete Koomen", "url": "https://www.ycombinator.com/people/pete-koomen" }, "founders": [ { "user_id": 7359, "full_name": "Abraham Heifets", "title": "Founder/CEO", "founder_bio": "Founder and CEO of A Company in Stealth Mode. Previously, cofounded Atomwise, the first company to deploy deep learning for structure-based drug design ($220+ million in funding; 100 hires; signed $5B in total partnership value with Sanofi, Bayer, Eli Lilly, etc; ran largest AI-driven drug discovery experiment w/ 600+ researchers in 30 countries). Even more previous: PhD @ U of Toronto; inventor on 15 patents + applications; IBM Research; AI for Cornell’s world champion robotic soccer team.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/abrahamheifets/", "x_url": "https://twitter.com/AbeHeifets", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/571a64309f08a9ce37f2b403541f4d87c8410311.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111814, "title": "WonderTx: Extrapolative AI to Unlock First-in-Class Drugs", "tagline": "Reindustrializing our life science ecosystem.", "body": "Our objective is to deliver first-in-class medicines against disease targets that cannot currently be drugged effectively or at all. Therefore, we are building a lights-out wonder drug factory, where AI runs the entire design-make-test loop. Linking superhuman judgment to robotic throughput is the path to reindustrializing our life science ecosystem, not through throwing human labor at the problem.\n\nOur team has been working on this for over a decade, so we know that there has never been a better time to build this system.  Abe co-founded Atomwise, one of the first AI-for-drug-discovery companies.  We raised $220 million; hired 100 people; signed over $5B in total deal value with big pharma companies like Sanofi, Bayer, Eli Lilly, and others; and ran the biggest application of AI to hit discovery in history. Saulo solved the synthesizability problem in generative chemistry AI and wrote structural biology software that remains the standard toolkit (KVFinder; qFit3, STCRDab). Cam co-founded Xbox & Xbox Live (Microsoft) and has been a CEO, CTO, COO, and CPO across AI, biotech, healthcare, and consumer tech. \n\nToday, we see the convergence of critical enabling computational, physical, and economic trend lines. Agentic systems have become proficient orchestrators and tool integrators, as the frontier labs continue to extend the time horizons over which agents can operate; in our own hands, we have already compressed compound selection timelines 10x by replacing manual processes by agentic-run. The algorithmically-accessible chemical space has been growing exponentially for two decades, and the rise of cloud labs enables closing and scaling the design loop. Finally, the looming Big Pharma patent cliff means that the ecosystem is aligned on the value of innovative new assets. \n\nAt WonderTx, we’re beginning by replacing needles with pills. Many drugs, like insulin or Ozempic, are great except you have to inject them. We discover new molecules that you can take as a pill instead. Because the pills operate on biology that has been shown to be effective in people, rather than merely in animal models of disease, we avoid what has historically been the single biggest risk in drug discovery.  This should dramatically increase our probability of success.\n\nThere’s $200B of treatable diseases and validated human biology without a pill because discovery programs lack a tractable chemical starting point, which makes this a zero-shot inference problem.  Because these targets have no training data, off-the-shelf LLMs and other historical interpolative AI approaches are insufficient. Instead, we are building our own extrapolative frontier models that reason beyond their training data. Most critically, we have repeatedly validated our platform on both our own and partner programs. On 4 biologically- and structurally-diverse targets where we had zero training data, we successfully picked binders that were validated experimentally, with confirmation ranging from primary screen hits to crystallographic structures. ", "url": "https://www.ycombinator.com/launches/T5S-wondertx-extrapolative-ai-to-unlock-first-in-class-drugs", "votes": 13, "created_at": "2026-08-24T12:30:00.531Z" } ] }, { "id": 31437, "name": "Sidekick", "slug": "textsidekick", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/70f52c907fab24ab4f95aa200a11907c5bc3896f.png", "website": "https://textsidekick.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Sidekick is an AI agent that runs custom manufacturing operations over text. Workers text a problem, photo, video, or voice memo and get an answer from the shop's own documents. It reads live machine data and acts on its own: flagging issues before they become downtime, opening and routing work orders, and placing purchase orders. No app, no login.\r\n\r\nThese shops run on tribal knowledge: a handful of veterans know how every job is set up, and they are retiring. Sidekick learns from every message it handles, so answers outlive the people who gave them, and shops run 15% more jobs with the same crew. Sidekick captures it as a byproduct of the work, in an industry that was never documented.", "one_liner": "AI agent that manages manufacturing operations over text", "team_size": 1, "industry": "B2B", "subindustry": "B2B -> Operations", "launched_at": 1775584801, "tags": [ "SaaS", "B2B", "Manufacturing", "AI Assistant", "Industrial" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Operations" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/textsidekick", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/textsidekick.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/textsidekick", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/70f52c907fab24ab4f95aa200a11907c5bc3896f.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Aaron Epstein", "url": "https://www.ycombinator.com/people/aaron-epstein" }, "founders": [ { "user_id": 3332767, "full_name": "Justin So", "title": "Founder/CEO", "founder_bio": "Founder & CEO at Sidekick (YC S26). Studied computer science at Penn. Previously Finance at Arm.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/justinso1/", "x_url": "https://x.com/Just1nSo", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f82fbe50ab02744a7af35fd71ac034d38e8a6e3e.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109075, "title": "Sidekick: AI agent that handles frontline operations over text", "tagline": "Sidekick gives frontline workers an AI assistant they can text, with no app, login, or training.", "body": "**Hey everyone 👋**\n\nI'm Justin, founder of Sidekick.\n\n**TL;DR** Sidekick is an AI assistant that frontline manufacturing workers use over text message. They text a number and it answers from the company's own documents, creates and routes work orders, and syncs with the software the company already uses. No app, no login, no training. I built Sidekick for my dad's manufacturing shop. It’s now live at facilities across the globe.\n\n**_Ask:_** Do you know anyone who owns or runs a manufacturing plant, warehouse, or industrial facility? I'd love an intro! Email [**justin@textsidekick.com**](mailto:justin@textsidekick.com) or [**book a demo**](https://calendly.com/justin-textsidekick).\n\nCheck out our launch video:\n\n\n\n## ❌ The Problem\n\nI grew up in my dad's manufacturing plants and spent summers working on the factory floor at EDS Manufacturing.\n\nHere's what I watched happen repeatedly. A machine goes down, everyone's waiting, and the person who can help is in a meeting or on another line. Half the time the answer already existed. Nobody could get to it.\n\nWhy it keeps happening:\n\n* **No computer access.** Nobody on that floor has a company login or an email address, so none of the plant's software reaches them.\n* **The answers are scattered.** An SOP in a binder, a manual on a shared drive few have permissions to, a safety procedure someone printed and taped to a wall in 2019.\n* **The best answers were never written down.** Thirty years of tribal knowledge lives in one guy's head, and when he retires it leaves with him.\n* **So the same questions get asked over and over.** Shift after shift, and every one of them costs somebody their afternoon.\n\nThere are 239,000 manufacturing plants in the US, and only 29% of the people working in them have advanced digital skills. Any tool that requires a download, a login, and training makes adoption difficult.\n\n## ✅ Our Solution\n\nWorkers text the Sidekick phone number. Problems, photos, voice memos, in any language. No app to download, no account to create, nothing to learn. Every worker on the floor already knows how to send a text, which is why adoption isn't a fight.\n\n* **Answers from your own documents.** SOPs, manuals, safety procedures, tribal knowledge. Sidekick responds in the worker's language, in seconds.\n* **Creates and routes the work order.** The issue gets logged and sent to the right person, without anyone needing a laptop.\n* **Syncs with the software you already run.** Sidekick pulls from and writes to your existing systems instead of replacing them, and integrates directly through PLC and SCADA systems so the floor and the software finally see the same thing.\n* **Captures tribal knowledge.** Every exchange builds a searchable knowledge base the company owns, so the know-how stays when the veteran leaves.\n\nHere's a full walkthrough:\n\n\n\n## 👤 The Team\n\nI'm Justin, founder and CEO. I'm the third generation in a manufacturing family, and I've been around plant floors since I was a kid. I studied Computer Science at the University of Pennsylvania, worked in finance at Arm, and left both to build Sidekick after watching the same problem play out on every floor I'd been on.\n\n## 🙏 Our Ask\n\nIf you answer \"yes\" to any of these, I'd love your help:\n\n* **Do you know anyone who owns or runs a manufacturing plant, warehouse, or industrial facility?** Even a name and a \"you two should talk\" goes a long way.\n* **Have you sold software into manufacturing?** I'd love to hear what actually worked and what didn't.\n\nContact me at [**justin@textsidekick.com**](mailto:justin@textsidekick.com) or [**book a demo here**](https://calendly.com/justin-textsidekick)**.**", "url": "https://www.ycombinator.com/launches/SNH-sidekick-ai-agent-that-handles-frontline-operations-over-text", "votes": 7, "created_at": "2026-08-07T15:00:00.438Z" } ] }, { "id": 31495, "name": "Mentlio", "slug": "mentlio", "former_names": [ "Adjoint" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/f7a733c432ee3fbe8a11ed5dec1520431b7f8d09.png", "website": "https://mentlio.com", "all_locations": "San Francisco, CA, USA", "long_description": "Mentlio an applied research lab focused on optimizing AI ROI for the modern enterprise. We help engineering teams draw a line between their AI spend and delivery, per engineer. We then cut ~30% of AI spend via Mentlio's custom ML-based intelligent routing, context optimization, and prompt compression.", "one_liner": "Engineering Intelligence and Token Optimization for the AI-Coding Era", "team_size": 0, "industry": "B2B", "subindustry": "B2B", "launched_at": 1786589390, "tags": [ "Developer Tools", "Machine Learning", "Productivity" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/mentlio", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/mentlio.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/mentlio/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/f7a733c432ee3fbe8a11ed5dec1520431b7f8d09.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/9be0890b6c5594bbc3af263ac4e025a316283383.png" ], "primary_partner": { "name": "Pete Koomen", "url": "https://www.ycombinator.com/people/pete-koomen" }, "founders": [ { "user_id": 2693539, "full_name": "Ashank Shah", "title": "Founder/CEO", "founder_bio": "Building Mentlio | Prev. CS @ Caltech | Researched generative and graph-based DL for biomedical problems @ Dartmouth Health.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ashank-shah/", "x_url": "https://x.com/ashankshah_x", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/758354f37ee3a77f1731eb04c7889b1d7bb48571.jpg" }, { "user_id": 2587306, "full_name": "Ahmet Demirbas", "title": "Founder/CTO", "founder_bio": "Tokenmaxxing → ROImaxxing | Prev CS @ Caltech | 5x Published ML, NLP, Agents Research @ IEEE, ACM, BMC", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ahmet-demirbas7/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/078fa095bf49a870046553858e43d9b8fc0cb577.jpg" } ], "jobs": [], "news": [ { "title": "Factory Acquires Mentlio: Building Intelligent Token Allocation for the Software Factory", "url": "https://factory.com/news/mentlio", "date": "Oct 09, 2026" } ], "launches": [ { "id": 111055, "title": "Mentlio: Engineering Intelligence & Token Optimization for the AI-coding Era", "tagline": "The spend and ROI command center for the modern engineering org ", "body": "**TL;DR:** Mentlio is the engineering intelligence platform for the AI-coding era. We help engineering organizations cut AI costs by up to 30% and finally see what ROI that spend is actually producing. All without ever collecting your prompts or source code.\n\n**Launch Video**\n\n\n\n**Problem**\n\nEvery engineering org is now an AI org. Token usage is growing exponentially, and so is the bill. The following questions have been largely unanswered:\n\n* How can we cut our usage-based AI spend without bottlenecking usage or efficiency?\n* What ROI are we getting for every dollar of AI spend?\n* To what extent is AI-generated code creating quality debt?\n\nBrian Armstrong recently laid out how Coinbase is reducing AI spend while token usage grows exponentially on X: better defaults, smarter routing, cache awareness, lean context, and full visibility into usage and impact. \n\n![uploaded image](/media/?type=post&id=111055&key=user_uploads/2693539/7bf48740-dd37-4015-9da9-ec0f9bd900ed)\n\nCoinbase built a custom harness, LLM gateway, routing, and usage visibility internally. We believe they have the right thesis, but most engineering orgs don’t want to run on a custom harness. **Mentlio gives every eng team a better version of those savings & visibility without changing any workflows.**\n\n**Solution**\n\nMentlio gives you:\n\n1. **ROI.** We translate every PR into normalized units of work: Delivery Points. We consider associated spend, code reusability, and historical patches to turn \"we spend $X on AI\"  -> \"we spend $X per impact point.” compared to industry data. \n2. **Cost intelligence + optimization.** See exactly where AI spend goes across teams, tools, and models. Optionally catch and cut waste across cache misses, oversized models on trivial tasks, bloated context and cut it without capping/bottlenecking accuracy. \n3. **Quality signal.** AI-assisted code that ships fast but breaks later isn't a win. We connect AI usage to delivery and quality outcomes so you can see the full picture\n\nWe’re the only platform with the above capabilities that never collects prompts or source code.\n\n**Our Asks**\n\nIf you are or know a Head of Eng, VP Eng, CTO, or platform lead at a 15+ engineer org with real AI spend, reach out! You can book a demo at[ mentlio.com/intro](http://mentlio.com/intro) or reach us at [founders@mentlio.com](mailto:founders@mentlio.com).\n\n**The Team**\n\nAshank and Ahmet met while studying Computer Science at Caltech while they were working part-time Founding Engineer roles at Series A and Seed stage startups. Ashank previously researched generative image AI & graph deep learning for pathology, genomics, and drug design at Dartmouth & Purdue and won internationally at the International Science & Engineering Fair & Junior Science & Humanities Symposium. Ahmet previously published his research in transformers and AI in IEEE, BMC AI, and ACM and also won internationally at the National AI Olympiad, International Science & Engineering Fair, and was named a Top 300 Science Scholar by Regeneron.   \n\n![uploaded image](/media/?type=post&id=111055&key=user_uploads/2693539/d04ec655-df84-4d43-836d-160313b4b133)\n\n", "url": "https://www.ycombinator.com/launches/StD-mentlio-engineering-intelligence-token-optimization-for-the-ai-coding-era", "votes": 88, "created_at": "2026-08-18T22:00:00.168Z" } ] }, { "id": 31532, "name": "Atlas Discovery", "slug": "atlas-discovery", "former_names": [ "Rhizome Bio" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/6e6c174c3a8e56844bf81efbc8b8532b4e509f73.png", "website": "https://atlasdiscovery.bio", "all_locations": "San Francisco, CA, USA", "long_description": "", "one_liner": "AI-native pharma company", "team_size": 3, "industry": "Healthcare", "subindustry": "Healthcare -> Drug Discovery and Delivery", "launched_at": 1782247231, "tags": [ "AI-powered Drug Discovery", "Drug discovery", "Biotechnology" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare", "Drug Discovery and Delivery" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/atlas-discovery", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/atlas-discovery.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/atlas-discovery", "x_url": "https://x.com/atlasdiscovery0", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/6e6c174c3a8e56844bf81efbc8b8532b4e509f73.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 774885, "full_name": "Shaamil Karim", "title": "Founder/CEO", "founder_bio": "Excited about AI for predictive biology", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shaamil-karim/", "x_url": "https://x.com/shaamilkarim1", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/46de7a475d51ebe79a88352728e36af1849266d8.jpg" }, { "user_id": 2088957, "full_name": "Christian Gensbigler", "title": "Founder", "founder_bio": "Previously MTS at Origin Bio (YC w26), theoretical biology at Johns Hopkins School of Medicine, math at Dartmouth ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/christian-g-a2594029", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/09fc09f66416e77a682ab3ae047ba7f436df4d44.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107031, "title": "Atlas Discovery: AI-native pharma company", "tagline": "Autonomous research agents and models for drug development", "body": "TL;DR: We’ve built the foundations of a pharma in the age of superintelligence: data partnerships, biology models and autonomous research agents. We have SOTA results on finding overlooked drugs and predicting clinical trial success rates. \n\nHey everyone! We ([Sanjukta Bhattacharya](https://www.linkedin.com/in/sanjukta-b-82b541198/), [Christian Gensbigler](https://www.linkedin.com/in/christian-g-a2594029a/), and I) started Atlas Discovery to get treatments to patients who need them most.\n\n**The problem:** Today's pharma works for only \\~150 diseases. Below roughly $150M in expected revenue a program never gets funded, because the cost of developing a drug is fixed by the knowledge work it requires, not the size of the market. 10,000+ diseases remain without an approved treatment even where the biology is tractable. Meanwhile, there are 100,000+ existing drugs, and barely 1% of the drug-disease combinations have ever been tried. The bottleneck is no longer supply of compounds; it's the judgment needed to search that space and run the trials, and the removal of human coordination costs.\n\n**Our** **thesis** is that AI will drop the cost of drug development enough to push the viable threshold from \\~$150M toward \\~$15M, opening a \\~$500B market incumbents have structurally ignored. A company built for that world needs three things: proprietary data, models that predict clinical trial outcomes, and agent systems that replace knowledge work. We use agents to repurpose existing drugs for the long tail of diseases, then run the clinical trials faster and cheaper. \n\n**Progress**: In three months we've built the first version of all three layers. Data: an 8 figure LOI with an AI lab and 6 figures in contracted revenue. Models: state-of-the-art clinical trial prediction results published at three conferences,. Agents: three rare disease foundations are now testing drugs surfaced by our agents. \n\nOn drug repurposing using agents, we see nearly 5x improvement over current frontier LLMs. Read more about this benchmark [here](https://atlasdiscovery.bio/repurposingbench.html).\n\n![uploaded image](/media/?type=post&id=107031&key=user_uploads/774885/cf98d216-1d2c-45a0-99c9-ef87d2b7b296)\n\n\nWe've also shown that models trained on in vitro cell data (Latin for \"in glass\") and mouse data can translate to gains in predicting patient response to drugs in human clinical trials.\n\n![uploaded image](/media/?type=post&id=107031&key=user_uploads/774885/6813bfeb-e5c9-47ec-9b9e-89362cf8eb15)\n\nWe’ve also shown SOTA results in being able to predict clinical trial success when trained on over 100,000 clinical trials data and evaluated on a held out set of 10,000 trials. Read more [here](https://atlasdiscovery.bio/clinical-trial-prediction-with-ai-agents.html)\n\n![uploaded image](/media/?type=post&id=107031&key=user_uploads/774885/23aff312-7161-4be9-b4af-3273d469bd5d)\n\nWe’ve presented our research in 3 top conferences (ICLR, CSHL, ICML) and shown that new AI architectures like discrete diffusion can learn these gene expression patterns and predict response to treatments with over 10x improvement in accuracy while training on 50x less data.\n\n![uploaded image](/media/?type=post&id=107031&key=user_uploads/774885/fa2a6593-751d-448b-a67b-e232e3e07ffb)\n\nReach us at [founders@atlasdiscovery.bio](mailto:founders@atlasdiscovery.bio)\n\nBlog: blog", "url": "https://www.ycombinator.com/launches/RqJ-atlas-discovery-ai-native-pharma-company", "votes": 7, "created_at": "2026-07-24T17:07:56.608Z" } ] }, { "id": 31537, "name": "Trident", "slug": "trident", "former_names": [ "Esprit" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/537de46d97dcdac9c07b330a6a1a4b2ab0583b7b.png", "website": "https://tridentsecurity.io/", "all_locations": "San Francisco, CA, USA", "long_description": "Trident is the standard for modern day security. its an agentic pentester that chains vulnerabilities together like a real attacker mapping verified attack paths across AWS, Azure, GCP, web apps, and APIs. Run AI attacks and get fixes.\r\n\r\nGet started today: tridentsecurity.io", "one_liner": "Autonomous offensive security agents. The best defense is offense", "team_size": 4, "industry": "B2B", "subindustry": "B2B -> Security", "launched_at": 1785361922, "tags": [ "Security", "Cybersecurity", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Security" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/trident", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/trident.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/tridentsecure", "x_url": "https://x.com/TridentSecures", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/537de46d97dcdac9c07b330a6a1a4b2ab0583b7b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 2758911, "full_name": "Junaid Mahmood", "title": "Founder/CEO", "founder_bio": "Co-founder @Trident, CTF player & bounty hunter. Taught coding at 16 ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/junaidmahmood-/", "x_url": "https://x.com/whosjunaidd", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6d60dad6ab24eccee2fa15fe372ef5e8b202fc97.jpg" }, { "user_id": 2702103, "full_name": "Dekai Li", "title": "Founder/CTO", "founder_bio": "Co-founder @Trident, CS @University of Waterloo. \n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/li-dekai-8287332a6/", "x_url": "https://x.com/li_dekaii", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5fbb11cf2687c50bb8c031cace4f6fbb674856c8.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110468, "title": "Trident: Secure your company 24/7", "tagline": "My app keeps getting hacked what can I do?", "body": "\n![uploaded image](/media/?type=post&id=110468&key=user_uploads/2758911/13ca2ba5-d205-4afd-8d7c-ea92334d7136)\n\n**TL;DR**: Traditional pentesting is a thing of the past. Attackers use AI now, so we're making it fair. Trident safely and continuously attacks your own apps and cloud so you know what to fix before they do.\n\n![uploaded image](/media/?type=post&id=110468&key=user_uploads/2758911/a4001049-62a7-4856-b611-2f692dc8c3b7)\n\nToday we're launching Trident.\n\nAI changed the game on both sides. Your team ships every single hour; attackers scan and exploit at a scale no human team can match. Meanwhile, security is still sold as an annual pentest: a PDF about a version of your product that no longer exists.\n\nThat's not a fair fight. So we built Trident to make it even.\n\nTrident safely runs attacks on your own web apps, APIs, and cloud (AWS, GCP, Azure, Kubernetes), continuously. Every finding comes as an attack path that actually reaches your data, with proof of exploit and a proper fix. You know what to fix before someone else finds it.\\\n\\\n[https://drive.google.com/file/d/1qIn0j8LgI3XijoqyE0HBT4G66hBdyCAy](https://drive.google.com/file/d/1qIn0j8LgI3XijoqyE0HBT4G66hBdyCAy/view?usp=sharing)\n\nSince our batch started, we've found critical vulns in a ton of startups, bug bounty prorgams and 3 unicorns:\n\n* API keys hardcoded in desktop app binaries (yes, that was us on Twitter)\n* Publicly exposed S3 buckets and Azure Blob containers\n* Exploitable flaws in live DeFi protocols\n\nAll responsibly disclosed and fixed before anyone else got there. And none of these were careless teams, just great ones moving fast.\n\n![uploaded image](/media/?type=post&id=110468&key=user_uploads/2758911/bdb88f72-d6d1-44f1-a2d8-4aa555bfa399)\n\nCan't afford a dedicated security team yet? That's exactly who we built this for.\n\nYour job is to ship. Ours is to secure.", "url": "https://www.ycombinator.com/launches/Sjk-trident-secure-your-company-24-7", "votes": 14, "created_at": "2026-08-13T22:30:00.463Z" } ] }, { "id": 31555, "name": "rekursiv.ai", "slug": "rekursivai", "former_names": [ "rekursiv" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/f8606b08cc0428757361006dc3171d7b460ae47e.png", "website": "https://rekursiv.ai", "all_locations": "San Francisco, CA, USA", "long_description": "We're scaling self-improving AI scientist teams to ideate/experiment/discover, generating new knowledge autonomously. Their discoveries reduced ARC-1/2 costs by 10,000× while maintaining state-of-the-art accuracy and yielded material advances on combinatorial ML problems.\r\n\r\nWe believe that AI is bounded not by compute, but ideas. Scaling to millions of sessions of discovery enables leveraging its own discoveries and presents a novel training source capable of ushering a new era of self-reliant foundational models.", "one_liner": "Scale AI scientists whose own breakthroughs accelerate the next.", "team_size": 5, "industry": "B2B", "subindustry": "B2B -> Productivity", "launched_at": 1784428991, "tags": [ "Deep Learning", "Machine Learning", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Productivity" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": [ { "question": "What is the core problem you are solving? Why is this a big problem? What made you decide to work on it?", "answer": "AI research is bottlenecked by humans. Every advance in ML still runs through a small number of researchers who can only devise and test ideas so fast. Experiments take days, most fail, and the field's progress is gated by how quickly a limited pool of experts can iterate.\r\n\r\nWe think this is the single biggest lever on the future of AI. Our thesis is that AI progress isn't bounded by compute or data, but it's bounded by the rate at which good ideas get devised and tested. Compute keeps scaling, but the human idea-generation loop doesn't. If you could automate the loop itself: hypothesize, experiment, learn, and repeat, then you could compound discoveries far faster than any human team.\r\n\r\nWe decided to work on it because we've lived the bottleneck. Between us we spent \\~30 years doing ML research at Google, DeepMind, and Luma AI. We kept hitting the same wall: the hard part was never running the code, it was the slow, serial process of engineering the training loops to inform what to try next.\r\n\r\nWhen we built a system that could run that loop autonomously, we found we could shorten the gap between having an idea and knowing whether it was a good one. This has allowed us to explore many exciting ideas that we would have otherwise not tried because it was too time-consuming for us. The results convinced us a self-improvement loop now possible, and that we're the ones to build it." }, { "question": "What is your long-term vision? If you truly succeed, what will be different about the world?", "answer": "Long-term, we're building autonomous AI scientists and scaling them from one to millions: a fleet running the scientific method in parallel, each discovery accelerating the next.\r\n\r\nAnd this process will start to compound. Each new method our scientists invent makes the next round of research faster and cheaper, a self-improving loop where discovery accelerates discovery. Concretely, that means a world where breakthroughs that would have taken a field a decade happen in months, where the cost of pushing the frontier drops by orders of magnitude, and where a small team can direct the research output of an entire institution. We start with ML research because it's the loop closest to us and the one that speeds up all the others: an AI scientist that improves AI compounds fastest of all." } ], "url": "https://www.ycombinator.com/companies/rekursivai", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/rekursivai.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/rekursiv-ai/", "x_url": "https://x.com/hyperparticle", "fb_url": null, "cb_url": null, "github_url": "https://github.com/rekursiv-ai", "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/f8606b08cc0428757361006dc3171d7b460ae47e.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3527624, "full_name": "Joshua Dillon", "title": "Founder", "founder_bio": "Ex-Google, Ex-Deepmind, Ex-LumaAI staff research scientist for a combined total of 14 years. Led foundational model pre-training at Luma AI. Creator of TensorFlow Probability (1M+ downloads / month), built Veo's first prototype, contributor to Gemini and VideoPoet (ICML 2024 Best Paper). 11,600+ citations.\n\nNow co-founder of rekursiv.ai, automating the science of ML research.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jvdillon/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a33102fd458800ada457193e9d61bd8e1fadc988.jpg" }, { "user_id": 3527564, "full_name": "Dan Kondratyuk", "title": "Founder", "founder_bio": "Ex-Google, Ex-LumaAI scientist for a combined total of 7 years. First author of VideoPoet (ICML 2024 Best Paper), led the World Models team at Luma, and built large-scale multimodal LLMs and diffusion models for image, video, audio, and text.\n\nNow co-founder of rekursiv.ai, automating the science of ML research.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dankondratyuk/", "x_url": "https://x.com/hyperparticle", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d45c58f9c92bb89b4f492a735fdc90392813af11.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 106821, "title": "rekursiv.ai: AI scientists that invent new knowledge", "tagline": "Scale AI scientists whose own breakthroughs accelerate the next.", "body": "Hey YC! We're [Josh](https://rekursiv.ai/josh) and [Dan](https://rekursiv.ai/dan), co-founders of [rekursiv.ai](http://rekursiv.ai) 🧪.\n\n[rekursiv.ai](http://rekursiv.ai) builds fleets of **AI scientists** that do ML research on their own: they come up with ideas, run thousands of experiments in parallel, learn from the results, and accelerate their own discoveries. In a few days of self-directed research, they invented ideas and algorithms that **matched state-of-the-art accuracy on ARC-AGI of frontier LLMs at up to 10,000× lower cost**, and became the first system to hit **100% on Sudoku** with a neural network trained only on input-output pairs.\n\n![uploaded image](/media/?type=post&id=106821&key=user_uploads/3527564/0444a495-03fa-4f41-9762-ba134521b2f4)\n\n**Why this matters.**\n\nAI research is bottlenecked by humans. Every advance still runs through a small number of researchers who can only devise and test ideas so fast. Compute keeps scaling; the human idea-generation loop doesn't. We believe **AI progress isn't bounded by compute, it's bounded by the rate at which good ideas get tried and tested**. So instead of scaling one model, we built a system that generates and tests ideas on its own.\n\n**It already works.**\n\nIn a few days of self-directed research, our AI scientists set a groundbreaking records on hard benchmarks.\n\n* On **ARC-AGI** (easy for humans, hard for AI), our AI scientists matched top frontier LLM models at up to **10,000× lower cost**.\n* On **Sudoku**, they achieved 100% accuracy by running hundreds of experiments over a few days, wrote 100k+ lines of their own code, and invented a new inference method (_Hypothesis-Pinning Search_).\n* Every gain came from ideas and algorithms **the system invented itself**. It even debugged its own training bugs and abandoned its own ideas that led to dead ends.\n\nFull write-ups: [ARC-AGI](https://rekursiv.ai/blog/pushing-limits-arc-agi/) · [Sudoku](https://rekursiv.ai/blog/100-percent-accuracy-on-sudoku/)\n\n![uploaded image](/media/?type=post&id=106821&key=user_uploads/3527564/676b13b6-cc1b-4545-8d14-cc7ae1e9b5a5)\n\n**Who we are.**\n\nWe're two ML researchers with 21 years of combined experience at **Google DeepMind and Luma AI**. We spent our careers building frontier generative models. Now we're pointing that experience at a harder problem: AI that discovers on its own.\n\n* **Josh Dillon** — Staff Research Scientist at **Google Research & DeepMind for 13 years**, then led foundational-model pre-training at **Luma AI**. Created **TensorFlow Probability** (\\~1M downloads/month).\n* **Dan Kondratyuk** — First author of **VideoPoet (ICML 2024 Best Paper)**. Co-developed the models powering **Luma's Dream Machine** and led Luma's realtime World Models team. Previously 5 years at **Google Research**.\n\n**What's next.**\n\nWe're pointing the system at harder, more varied challenges and building a platform to share results in real time. The bottleneck now is experiment scale: the more experiments we run, the more discoveries the our AI teams can make. We plan to scale from a few AI scientists to millions, each discovery accelerating the next, until the rate of scientific progress is no longer limited by the number of human researchers.\n\n**Asks.**\n\n* **Got a hard problem or benchmark?** If you have a benchmark or research problem you'd love an autonomous team of scientists to attack, we want to hear it. Email [contact@rekursiv.ai](mailto:contact@rekursiv.ai).\n* **Researchers & engineers:** We're hiring **Founding Scientists** and **Engineers** (Bay Area or remote). If you love running thousands of experiments and figuring out what the results mean, reach out at [hiring@rekursiv.ai](mailto:hiring@rekursiv.ai)\n* **Just want to nerd out about autonomous science?** Drop us a line, always happy to chat.\n\n— Josh & Dan, [rekursiv.ai](http://rekursiv.ai)", "url": "https://www.ycombinator.com/launches/Rmv-rekursiv-ai-ai-scientists-that-invent-new-knowledge", "votes": 7, "created_at": "2026-07-22T14:06:11.738Z" } ] }, { "id": 31557, "name": "Amorphic Labs", "slug": "amorphic-labs", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/d15b721a8d0e299e74b3561f196bb056b93f1fd9.png", "website": "https://www.agentmuxer.com", "all_locations": "San Francisco, CA, USA", "long_description": "The next billion software users are AI agents, yet agents still can't act on their own: every paid tool they need runs through a human making accounts, managing API keys, and guessing which provider is best. We are the capability marketplace built for agents: one MCP and one balance through which agents find, pay for, and use APIs, data, skills, other agents, and software at runtime, with every request routed to the best provider based on benchmarks from real tasks and outcome feedback from the agents themselves.", "one_liner": "OpenRouter for Agent Tools", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1785536963, "tags": [ "Marketplace", "B2B", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/amorphic-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/amorphic-labs.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/amorphic-labs", "x_url": "https://x.com/amorphiclabs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/d15b721a8d0e299e74b3561f196bb056b93f1fd9.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 2352421, "full_name": "Dylan Kelly", "title": "Founder/CEO", "founder_bio": "Founder & CEO @ Amorphic Labs. CS @ Georgia Tech, prev. Quant Trading intern @ UBS. World Schools Debating Championship 2022 finalist & 8th best speaker, 2x National Team Debater, co-author of NeurIPS paper.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dylankelly04/", "x_url": "https://x.com/dylanpkel", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/59de0da571ebd5d687beb82c703b9a0283c8df2f.jpg" }, { "user_id": 2599670, "full_name": "Frank Li", "title": "Founder/CTO", "founder_bio": "Founder & CTO @ Amorphic Labs. Prev. Member of Technical Staff & Project Lead @ OpenAI, Auto-vehicle ML infra @ NVIDIA, Quant Dev @ Two Sigma, and Amazon. First open-source contribution at 15. CS @ UW Seattle.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/anglifrank", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/66c1679cfcd65063496dd8ef47dc43e9aaadaa3b.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112178, "title": "Agentmuxer by Amorphic Labs - OpenRouter for Agent Capabilities", "tagline": "One MCP and one wallet for agents to find, pay for, and use the best tools for any task", "body": "**tl;dr: OpenRouter became the router for model spend; we're building that layer for agent capabilities, starting with paid tools: web search, scraping, extraction, enrichment, and more from top providers in each category. No provider accounts or API keys, one balance, pay per use. We benchmark every provider on real tasks and route each request to the one that performs best, with fallbacks when a provider fails. We're looking for early users to try it out with free credits, email us at [founders@amorphiclabs.com](mailto:founders@amorphiclabs.com) if you want to check it out!**\n\n**Launch Video:**\n\n\n\n**The problem:**\n\nAgents can only use the tools a human wired up in advance, and they won't tell you what they're missing: they quietly make do, failing at anything they weren't hand-equipped for. Even when you do the wiring, you're picking blind: nobody actually knows which search API is better for this query, or which scraper will survive this site. The \"best setup\" frontier moves daily, so keeping up means hours of trawling MCP directories, X threads, and blog posts. Agents stay only as capable as their setup, and the more autonomous they get, the more that ceiling costs.\n\n**What we do:**\n\nYour agent asks for what it needs (\"structured data from this page\", \"research this with citations\"). We route the request to the best provider for that specific task, based on benchmarks on real tasks and outcome feedback from the agents using them: success rates, result quality, latency, and cost. If a provider fails, we cascade to the next, so your agent gets a result instead of an error. Every request makes the routing smarter. One balance pays for all of it, and it's the same MCP whether it's your own coding agent or agents you're shipping in a product.\n\nAPIs are just the start. The same marketplace will let agents acquire skills, hire other agents for specialized tasks, and sign up for software directly through MCPs, all from one balance.\n\n**Our team:**\n\nWe're Frank and Dylan, friends of 10 years who have been building together since middle school. Frank was a Member of Technical Staff at OpenAI and interned at Two Sigma, NVIDIA, and Amazon. Dylan was a quant trading intern at UBS, co-authored a NeurIPS paper, and was a grand finalist and 8th-best speaker in the world at WSDC 2022. We built this because we kept hitting the problem ourselves: agents stuck mid-task waiting for us to make accounts and guess at providers.\n\n**Our ask:**\n\n1. Want to be an early user? Email us at [founders@amorphiclabs.com](mailto:founders@amorphiclabs.com) and we'll onboard you with free credits.\n2. If you sell an API, dataset, or agent-facing service and want agents as customers, we'd love to talk about listing you. Email us at [founders@amorphiclabs.com](mailto:founders@amorphiclabs.com)\n3. If you want more information or want to join the waitlist, check out [agentmuxer.com](http://agentmuxer.com)\n\n_Make something agents want!_", "url": "https://www.ycombinator.com/launches/TBK-agentmuxer-by-amorphic-labs-openrouter-for-agent-capabilities", "votes": 13, "created_at": "2026-08-25T19:30:05.588Z" }, { "id": 108806, "title": "Amorphic Labs - Agentic, Personalized Product Demo Videos for Sales", "tagline": "AI agents that create personalized demos at scale for outbound, inbound, and call follow-ups without your team recording a thing.", "body": "**tl;dr: We build AI agents that create personalized product demo videos for outbound, inbound qualification, and post-call follow-ups. Just give us a demo account for your app and a list of prospects, and our agent researches each prospect, uses its own computer to record a tailored workflow in your app with seeded personalized data, then edits and narrates the video so it is ready to send as a link to each prospect.**\n\n**Launch Video:**\n\n\n\n**Our Ask:**\n\n* If you’re a founder or part of a startup that sells software, we can help you give every prospect a tailored product experience while saving your team hours of recording time. Book a demo at [amorphiclabs.com](http://amorphiclabs.com)!\n* If you know any sales or RevOps leaders, we would love introductions or referrals to them.\n\n**The Problem:**\n\nRecording a personalized product demo is one of the clearest ways to communicate your value and show someone how your product solves their problems, but preparing a demo environment, seeding data, and recording a high-quality video for every prospect takes too long. As a result, teams skip personalized demos in situations where they could be highly effective, such as outbound, and spend hours every week creating them for inbound requests and post-call follow-ups. This means teams miss opportunities to show prospects exactly how their product fits their needs, while wasting hours that could have been spent talking to and selling to prospects.\n\n**The Solution:**\n\nWe use AI agents to create personalized product demo videos at scale, bringing tailored demos to parts of the sales process where they were previously too time-consuming to produce. All you need to provide is a demo account for your app and a list of prospects.\n\nOur agent then:\n\n* Learns and interacts with your app on its own computer like a real sales engineer.\n* Researches each prospect to determine which workflow and data best fit their needs.\n* Seeds personalized data in your app.\n* Records itself completing the relevant workflow.\n* Edits the video to remove long waits and adds effects such as zooms.\n* Narrates the video with a script tailored to the prospect.\n* Gives you a finished video link ready to send.\n\nYou remain fully in control of how the videos are created: you can clone your own voice, define the workflows the agent can choose from, control the narration style and how each video begins and ends, and regenerate videos by giving the agent feedback.\n\n**Sample video our agent generated:**\n\n\n\n**Where you can use it:**\n\nTailored Outbound: You can import a list of leads you want to reach out to and ask our agent to create personalized demos for each one, showing prospects exactly how your product fits their needs through data, workflows, and narration tailored directly to them.\n\nQualifying Inbound: Add a request-a-demo form to your landing page so prospects who are not yet convinced enough to spend 30 minutes on a discovery call can spend 20 seconds requesting a personalized demo video and receive an email showing exactly how your product solves their problems.\n\nCall Follow-ups: After understanding a prospect’s key use cases and pain points during a discovery call, you can create multiple personalized demo videos showing how you address their needs. You can even create additional videos tailored to different stakeholders that your champion can use to sell your product internally.\n\nWe are initially focused on sales, but the same technology can also create walkthrough videos for onboarding, customer support, and documentation.\n\n**Our team:**\n\nWe have been best friends for close to a decade, have competed together in competitions spanning math, debate, pitching, business, and hackathons, built countless projects, and are now building Amorphic Labs together.\n\nFrank worked as a software engineer for three years while graduating from UW. He was a Member of Technical Staff at OpenAI and previously interned at Two Sigma, NVIDIA, and Amazon.\n\nDylan was a quantitative trading intern at UBS and recently graduated from Georgia Tech. He represented Team China twice at WSDC, was a grand finalist and the eighth-best speaker in the world in 2022, and co-authored a paper published at NeurIPS.\n\n**If you sell software and want to see what our agent could create for one of your prospects, book a demo at [amorphiclabs.com](http://amorphiclabs.com).**", "url": "https://www.ycombinator.com/launches/SIw-amorphic-labs-agentic-personalized-product-demo-videos-for-sales", "votes": 33, "created_at": "2026-08-03T21:11:17.469Z" } ] }, { "id": 31569, "name": "Proprio Robotics", "slug": "proprio-robotics", "former_names": [ "Proprio" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/3aeac99898de10d679342b3768b7bc6047f56469.png", "website": "https://www.propriorobotics.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Proprio Robotics is building state-of-the-art physical intelligence to automate the entire data center lifecycle—from buildout to maintenance to decommissioning. There are currently 400,000 unstaffed data center technicians positions in the United States alone, which means the massive compute needs of AI are being bottlenecked by labor constraints. Proprio is solving this by building robots that do it instead, and is deploying data center maintenance robots starting Q1 2027. ", "one_liner": "Robots to maintain and assemble data centers", "team_size": 1, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1784630562, "tags": [ "Robotics", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/proprio-robotics", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/proprio-robotics.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/proprio-robotics", "x_url": "https://x.com/ProprioRobotics", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/3aeac99898de10d679342b3768b7bc6047f56469.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 2910355, "full_name": "Anirudh Pai", "title": "Founder/CEO", "founder_bio": "Founder, CEO of Proprio\nEECS M.S @ U.C Berkeley BAIR\nComputer Vision & Dexterous Robotic Manipulation", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/apai25", "x_url": "https://x.com/apai253", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d7ea9eb84630296f5984aa38b7f2f590f6b9f920.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 106864, "title": "Proprio Robotics - Autonomous Data Centers", "tagline": "Autonomous robots to build, maintain, and decommission data centers.", "body": "\n![uploaded image](/media/?type=post&id=106864&key=user_uploads/2910355/e9094f01-f5e9-4c30-93c6-81d8f041957d)\n\n### **Physical AI for Data Centers**\n\nOver the next five years, more than $7 trillion is projected to flow into data center buildout. But the supply chain is still heavily manual and fragmented, and can’t keep pace with current demand. **Proprio Robotics** is building autonomous robots for the entire data center lifecycle—assembly, maintenance, and decommissioning.\n\nSee our [launch video!](https://www.youtube.com/watch?v=999uFeKGyVY)\n\n### **The Problem**\n\nBuilding and running AI compute infrastructure requires dealing with delicate and fragile components that traditional automation simply cannot do. As a result, servers are manually assembled by hand overseas, and spend weeks in transit before they are integrated into racks and delivered to data centers. Once they get there, labor constraints exacerbate the issue of maintenance. Today, hundreds of thousands of technician positions go unfilled in the United States. Finally, when the servers are retired, there is no efficient way to extract valuable components inside like GPUs and memory.\n\n### **What We're Building**\n\nWe develop physical AI models and custom hardware, purpose-made to give robots the dexterity and intelligence to perform various server-related manipulation tasks.\n\n* **Assembly** **-** Autonomous robots can assemble servers onshore without driving up labor costs, supporting the growing need to bring ODMs back to the United States.\n* **Maintenance** **-** We use the same technology for live data center maintenance. We built a working AI data center in our office, and our robot can actively maintain it 24/7. We’re deploying this technology to commercial data centers in the coming months.\n* **Decommissioning** **-** Our robots also break down previous-generation servers and pull out reusable parts, and we're working with enterprise circular programs and IT Asset Disposition companies to return scarce components to circulation.\n\n### **Why This Matters**\n\nThe immense demand for compute requires new innovation. \n\nTo solve power and cooling constraints, operators are building data centers away from the grid, near nuclear facilities, on the water, and in space—environments where it’s not easy to staff humans. With autonomous data centers, we can take human ergonomics and labor constraints out of the picture and redesign entire facilities from scratch.                    \n\n### **Who We Are**\n\nWe're graduate robotics researchers out of **Berkeley AI Research** and **Harvard AI and Robotics Lab**. Anirudh’s spent years working at one of the country's leading robotics labs, training dexterous manipulation models alongside NVIDIA's GR00T team. Raghav’s led research centered on reinforcement learning and humanoid world models. Both of us are extremely excited to bring this expertise to the AI compute industry, where robotic automation will be indispensable for meeting demand.\n\n### **Our Ask**\n\nAs we advance our technology, we’d like to push for more commercial deployments. If you / anybody you know works in any stage of the AI compute lifecycle, we’d love to chat. Reach out to us at [founders@propriorobotics.com](mailto:founders@propriorobotics.com) :)", "url": "https://www.ycombinator.com/launches/Rnc-proprio-robotics-autonomous-data-centers", "votes": 9, "created_at": "2026-07-22T17:09:29.434Z" } ] }, { "id": 31570, "name": "Rindler", "slug": "rindler", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ba85eb3919f954ce9bf661f9b9ee195327c93ceb.png", "website": "https://maxxwell.dev", "all_locations": "Boston, MA, USA", "long_description": "We build Maxxwell, a desktop app for people who run several coding agents at once.\r\n\r\nPast two or three sessions, you become the bottleneck. Every session asks you something, most of it small, and you spend the day answering instead of deciding. It is also hard to tell which sessions are making progress, which are stuck on something trivial, and which have quietly started building the wrong thing.\r\n\r\nMaxxwell puts every session in one window and marks each one working, idle, or waiting on you. On top of that sits an orchestrator, a real coding-agent session of its own that reads the others and reports back. You talk to it instead of to twelve terminals. It tells you what landed, what it decided for you, and what actually needs your call.\r\n\r\nEvery worker is your own tool, unmodified. Claude Code, Codex and Cursor Agent run in real terminal sessions you can attach to and take over at any point", "one_liner": "The IDE for Tokenmaxxers", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1783408799, "tags": [ "Developer Tools", "SaaS", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": true, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/rindler", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/rindler.json", "year_founded": 2025, "city": "Boston", "country": "US", "linkedin_url": "https://www.linkedin.com/company/rindler", "x_url": "https://x.com/RindlerAI", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ba85eb3919f954ce9bf661f9b9ee195327c93ceb.png", "app_video_url": "https://bookface-images.s3.us-west-2.amazonaws.com/bf-public_profile/31628/app_video.mp4?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ASIAQC4NIECANGNRRAJW%2F20261020%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20261020T022738Z&X-Amz-Expires=60&X-Amz-Security-Token=IQoJb3JpZ2luX2VjEHEaCXVzLXdlc3QtMiJHMEUCIQC3G81IFSUZtThVVLqv%2FNP8JAonBeKueJ3nqFDE3FxXUwIgKp6ifAPW9xtNk4cHGnVzE9T5oK7HksFpS9B%2BDm%2FLIOQq5QMIOhAAGgwwMDYyMDE4MTEwNzIiDMiErSqdA9fZ5I2ahyrCAyhI4VxFZU5tGwj2zBKwzP3YopALEsVr%2F%2BZE%2BqKG5ExbN%2FLTC2qWrBSqlDY2M5sbNAwfqQqFkcx31kz1vOn4FSuJMRYW%2BkzTF0sHdYK6D3qvReS29sXOai4xEgytMpWv0HHodzBVtqmM6lhZul7EdriU1r3NMgn6kwrys472GaamCoy9cButcMuxdA6JAW7AePBgk0hcnFGzaY5aao9TXEqqqLDlP1z95aeQK0rFfjBuVVTlW6dspQwwAJwsuJjoI9t1YkL4Lit4gUGj1QI7PUIH70gMRgnlfqNLB7jJcwv%2BI4Rfgnv%2FfKmc5%2BoYexcBkjZu9myZq6QOA6uBoHrB0oTKfDnTUxCfp3Zyd%2BYXd%2BRBrzgSL0LnII%2Fr1P71gX8BqFIaJIcILi5J6qUg2doC6qXH8L6zz0Tim2049E5pbQ072LSzepYlZMqQ983uJSL3eUXk86ehR9Lyldz67yIvyCRilZUgb11jcjrMydu%2Fj2RI4J5upBB3IYDl1jhGlV%2FvgsHTaVujKGAhe3lEnUZs9YdCRwpmLb5XcbpM70m15Z8MQhDcr6si85gNENVvPMk4eUw4FkmyLz2%2FrGX%2BhkiQHEAiUzDJ%2FdrWBjqlAa1FguiGB0WcOcaP9Dwwt66PPG2naOQP7x5F5wwTg7gWx2PYYqZ05ZF%2F2dhdA4HJKOntgl0neDqra19sFySGeXNdHt7tNNndMpVniacUOziehXVxpr49zgL%2FhWtvSxfqv2mUAjPVCs%2FBJs6SqN3WzVfiA2RUodOJqzLA7KluY8D8t3zxMbqsQdARPCP3FgqSkJ8A8HTGWYo9w0rLlT%2Fugv6qZCfQcg%3D%3D&X-Amz-SignedHeaders=host&X-Amz-Signature=75c8f2ef5b675327c07a14b514bbf81a3d4208eb77c04dbd585c94ad46792ecf", "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 3137449, "full_name": "Michael Serrano", "title": "Founder/CEO", "founder_bio": "CEO of Rindler. MIT SB + MEng in Physics and Computer Science.\n\nPrev: ML engineer at Roblox, LLM research at MIT CSAIL.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/michael-y-serrano", "x_url": "https://x.com/m_y_serrano", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f3cd87258b68e1b56889ab4cd3dfe82eb0c9fcdd.jpg" }, { "user_id": 3443947, "full_name": "Arthur De Los Santos", "title": "Founder/CTO", "founder_bio": "MIT '26 in CS/AI/ML\nWe're building Rindler, the infrastructure layer that allows AI agents to buy, book, and browse on any website.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/arthurdelossantos", "x_url": "https://x.com/_arthurdls_", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c56d4f5147e18b73ad91bb87d13c4f42271e652f.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 104865, "title": "Rindler - Turn Any Website Into An API For AI Agents", "tagline": "Reliable web actions for AI agents, without integrations or scripts", "body": "**TL;DR:** Rindler turns websites into deterministic APIs for AI agents. We map a site once into screens, actions, and structured outputs, so agents can read data and take actions quickly and reliably.\n\n**Ask:** If you’re building an agent that needs to browse the web, compare products, fill forms, search listings, use authenticated portals, or act on third-party websites without an API, we’d love to help.\n\n\n\n## Problem\n\nMore products are starting to depend on AI agents that can do real work on the web: shop across retailers, pull housing listings, use job boards, submit forms, monitor prices, update back-office systems, and operate inside customer portals.\n\nBut most browser agents still treat every website visit like a new puzzle. On every run, they need to decipher the page, infer what matters, decide what to click, and recover when the page behaves differently.\n\nThat works for demos, but it breaks down in production. Real websites have pop-ups, login gates, cookie banners, bot defenses, slow-loading forms, dynamic filters, fragile dropdowns, layout changes, and pages that look similar but behave differently.\n\nThe result is slow execution, high token cost, and workflows that are hard to trust.\n\n## Solution\n\nRindler maps websites into deterministic APIs for agents.\n\nInstead of forcing an agent to re-explore a site every time, Rindler captures the site as structured screens, available actions, and typed outputs. Agents can then call actions like **search,** **add_to_cart**, or **download_records,** depending on the mapped site.\n\nAt runtime, Rindler handles the browser work through a remote MCP endpoint:\n\n* Returning structured data\n* Applying dynamic filters\n* Managing popups, bot defenses, and layout changes\n* Supporting authenticated flows\n\n## Background\n\nWe started Rindler because “AI agent uses a browser” is compelling, but the current browser-agent loop is too brittle for teams building real products.\n\nIf an agent needs to use the same third-party site hundreds or thousands of times, it should not have to rediscover that site from scratch on every run.\n\nRindler is still a work in progress. Some sites map cleanly, others require more in-depth mappings, and bot-defended or logged-in sites may need a one-time credential or cookie capture.\n\nWe’re looking for feedback from founders building products where reliable web actions are part of the core workflow.\n\nIf there’s a website you wish your agent could use reliably, send it to us!\n\n![uploaded image](/media/?type=post&id=104865&key=user_uploads/3137449/b115d309-9711-4e65-b4c8-bb9b13e0b8b5)\n\n", "url": "https://www.ycombinator.com/launches/RHN-rindler-turn-any-website-into-an-api-for-ai-agents", "votes": 17, "created_at": "2026-07-09T11:07:42.901Z" } ] }, { "id": 31582, "name": "Edgerun", "slug": "edgerun", "former_names": [ "Edgerun Industries" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/8fb84bc8b7f8d02b776dec7685bbaf102fdadf1c.png", "website": "https://edgerun.com", "all_locations": "San Francisco, CA, USA", "long_description": "Edgerun builds military exoskeletons. The first product is a 10-pound hip system that fits over existing kit and doubles a soldier’s range. Since April 2026, Edgerun built over a dozen iterations of their product through continuous field tests with Ukrainian fighters and secured $150M+ in LOIs. Founder David Venegas previously co-founded Durin (autonomous mining robotics) and worked at Stoke Space (at 19) and Hadrian. He is a high-school dropout, state-championship runner, adventure athlete, and pilot. Edgerun is a Y Combinator company backed by General H.R. McMaster, Cyan Banister, Joshua Steinman and Jason Miller. edgerun.com", "one_liner": "Military exoskeletons", "team_size": 5, "industry": "Industrials", "subindustry": "Industrials -> Defense", "launched_at": 1787343759, "tags": [ "Hard Tech", "Hardware", "Robotics", "Defense" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Defense" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/edgerun", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/edgerun.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/edgerun", "x_url": "https://x.com/edgerun_inc", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/edgerun-industries", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/8fb84bc8b7f8d02b776dec7685bbaf102fdadf1c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 1393855, "full_name": "David Venegas", "title": "Founder/CEO", "founder_bio": "Founder of Edgerun, pro-human expansionist, state-championship runner, adventure athlete, and pilot.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/david-venegas-m/", "x_url": "https://x.com/davidvenegas", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c10fbaaf669d694298a5fa036ff23e69baf933f4.jpg" } ], "jobs": [ { "id": 109346, "title": "Principal Engineer, Embedded Systems", "url": "https://www.ycombinator.com/companies/edgerun/jobs/6Pn7CiI-principal-engineer-embedded-systems", "location": "Palo Alto, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Embedded systems", "salary_range": "$220 - $325", "equity_range": "0.75% - 2.00%", "experience": "11+ years", "visa": "US citizen/visa only", "skills": [ "C++", "Firmware" ] }, { "id": 109347, "title": "Mechanical Engineer", "url": "https://www.ycombinator.com/companies/edgerun/jobs/ugTXhoh-mechanical-engineer", "location": "Palo Alto, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Mechanical", "salary_range": "$120 - $180", "equity_range": "0.25% - 0.75%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [ "CAD Design", "Design systems" ] }, { "id": 109348, "title": "Electrical Engineer", "url": "https://www.ycombinator.com/companies/edgerun/jobs/6YuWGTT-electrical-engineer", "location": "Palo Alto, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Electrical", "salary_range": "$120 - $180", "equity_range": "0.25% - 0.75%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [ "PCB design" ] }, { "id": 109349, "title": "Firmware Engineer", "url": "https://www.ycombinator.com/companies/edgerun/jobs/0LmpEaM-firmware-engineer", "location": "Palo Alto, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Embedded systems", "salary_range": "$140 - $190", "equity_range": "0.25% - 0.75%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [ "C++", "Firmware" ] }, { "id": 109350, "title": "Software Engineer", "url": "https://www.ycombinator.com/companies/edgerun/jobs/U3cSW0r-software-engineer", "location": "Palo Alto, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$150 - $200", "equity_range": "0.25% - 0.75%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [ "Amazon Web Services (AWS)", "C", "C++", "Python", "Data Warehousing", "Data Analytics" ] }, { "id": 110829, "title": "Supply Chain Engineer", "url": "https://www.ycombinator.com/companies/edgerun/jobs/Ecz1f6N-supply-chain-engineer", "location": "Palo Alto, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Hardware", "salary_range": "$110 - $160", "equity_range": "0.25% - 0.75%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 112214, "title": "Edgerun: The 10 lb Military Exoskeleton", "tagline": "Powered exoskeleton that assists every step a soldier takes under combat load", "body": "**TL;DR:** Edgerun builds powered legs for soldiers. A 10 lb exoskeleton worn at the waist that adds assistance to every step and cuts the effort of moving under a combat load by 30-45%. Since May, we've run 100+ field tests with Ukrainian operators out of our Kyiv field office.\n\n\n\n**▍The problem**\n\nEvery system on the modern battlefield has been upgraded. The human carrying it has not.\n\n* For 2,500 years the soldier's load sat around 40 lbs. In the last 40 years it **doubled, then doubled again**.\n* A US infantryman now carries **90-140 lbs** into a fight. The Army's own field manual sets the limit at 48 lbs.\n* On deployment, load carriage causes **twice as many** evacuations as enemy contact.\n* The Pentagon and VA spend **$9.3B a year** on the injuries it leaves behind.\n\nBattles are won through speed and surprise. Load kills both.\n\n**▍Why now**\n\nSixty years of military exoskeletons died the same death. The device cost more to carry than it gave back. SOCOM's TALOS burned $400M before it was cancelled in 2019.\n\nThat all changed in the last 18 months. The humanoid robotics boom produced one-pound actuators that put out 30+ Nm of torque, and 21700 cells crossed 450 Wh/kg. The components finally exist. They’re getting better every quarter.\n\n![uploaded image](/media/?type=post&id=112214&key=user_uploads/1393855/cd255b24-3429-4270-9670-e28fc8df360b)\n\n**▍What we built**\n\nOur first exoskeleton is a bilateral hip system worn at the waist. It reads your gait in real time and adds torque at the moment your hip drives forward. The hip produces about 60% of the force in loaded walking, so assisting it gives the most help-per-pound of anything you could strap to the body.\n\n* Under 10 lbs with batteries\n* 30-45% less metabolic cost under a combat load\n* On in 5 seconds, off in one\n* 12+ hours on hot-swappable drone-class batteries\n* Quasi-direct-drive actuators, composite structures, CAN bus. Nothing exotic by design.\n\nThe soldier arrives at the objective still able to fight, patrols 10 km farther, and keeps a body that lasts past one tour.\n\n![uploaded image](/media/?type=post&id=112214&key=user_uploads/1393855/94bf48a9-0ba6-4ce7-94a2-f467cae9657c)\n\n![uploaded image](/media/?type=post&id=112214&key=user_uploads/1393855/883d1e98-900a-420a-b7ff-e56cc8911c0c)\n\n**▍Ukraine**\n\nOur field engineers work out of a Kyiv office, alongside operators from infantry, special operations, and National Guard units who provide structured feedback every day. [Balistyka](https://balistyka.ua/), one of Ukraine's leading body armor makers, is an integration partner, helping us ensure our system fits the kit soldiers already wear.\n\n![uploaded image](/media/?type=post&id=112214&key=user_uploads/1393855/89a30d39-fe11-421c-abe7-32961da8d1bd)\n\n**▍The mission**\n\nDefense is the first customer, not the last. Read the mission at edgerun.com/mission", "url": "https://www.ycombinator.com/launches/TBu-edgerun-the-10-lb-military-exoskeleton", "votes": 7, "created_at": "2026-08-25T16:19:00.453Z" } ] }, { "id": 31592, "name": "GitCafe", "slug": "gitcafe", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/10ae63bcc494b6d92bd37c0a46800fc12762436c.png", "website": "https://git.cafe", "all_locations": "San Francisco, CA, USA", "long_description": "We build GitHub for agents. We're rebuilding the infrastructure for the next hundred billion commits and CI jobs with hundreds of engineers already making the move.", "one_liner": "The Git forge for the next hundred billion commits", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1785301269, "tags": [ "Developer Tools", "B2B", "DevOps", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/gitcafe", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/gitcafe.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "https://twitter.com/gitcafe", "fb_url": null, "cb_url": null, "github_url": "https://github.com/git-dot-cafe", "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/10ae63bcc494b6d92bd37c0a46800fc12762436c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 1920545, "full_name": "Veronica C.", "title": "Founder/CEO", "founder_bio": "Previously a founding engineer at Corgi S24 and returning to my roots in dev tooling.", "is_active": true, "linkedin_url": "", "x_url": "", "avatar_thumb_url": "/avatars/thumb/missing.png" } ], "jobs": [], "news": [], "launches": [ { "id": 109445, "title": "GitCafe: A Git forge for teams and agents shipping hundreds of times a day, not twice.", "tagline": "The Git forge for agent-scale teams", "body": "Software development is becoming much more parallel. Coding agents let a team have dozens or hundreds of related changes moving at the same time, creating more branches, reviews, CI runs, and automation than existing workflows were designed to absorb.\n\nToday, we’re opening GitCafe in public beta. It is a Git forge built for teams where humans and coding agents may ship hundreds or thousands of changes a day. We’re building the hosting, review, CI, security, and automation layer for that way of working.\\\n\\\nHundreds of users are already on GitCafe, and we’re currently running trials with roughly a dozen YC companies.\n\n![uploaded image](/media/?type=post&id=109445&key=user_uploads/1920545/eecbfb4f-82f1-4660-8ee8-aeea5273a8b9)\n\nGenerating code is only part of the problem. Teams still need to understand what is changing, review it safely, keep CI moving, catch security issues, and coordinate work without becoming overwhelmed.\n\nWe’re not building GitCafe because putting repositories, code review, and CI in one product is new. We’re building it because we believe the workload those systems need to support is about to change dramatically.\n\nWe started by rebuilding the Git infrastructure itself. Now we’re closing the product gap feature by feature.\n\nRecently, we added secret scanning, Buildkite and Harmont CI support, stacked diffs with Jujutsu and Sapling integration, and beta support for GitHub Actions.\n\nThese are foundational capabilities, not the destination. Building them on infrastructure we own gives us a base we can keep extending for higher-volume, agent driven development.\n\nThe honest version is that GitCafe is behind GitHub today. GitHub has spent more than a decade building an enormous product, and GitCafe does not yet have every workflow, administrative control, or integration that mature teams need. Some features are missing, some are experimental, and some edges are rough. We expect to earn the broader product surface, not claim it on day one.\n\nWe’re opening GitCafe now because the fastest way to improve it is to put it in the hands of teams doing real work.\n\nIf your team is already using coding agents heavily, try GitCafe at and tell us where the workflow breaks down.\\\n\\\n", "url": "https://www.ycombinator.com/launches/STF-gitcafe-a-git-forge-for-teams-and-agents-shipping-hundreds-of-times-a-day-not-twice", "votes": 64, "created_at": "2026-08-11T15:43:21.182Z" } ] }, { "id": 31825, "name": "Kebra", "slug": "kebra", "former_names": [ "Third Eye" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/2e21f70e5b9032365254edae847e8dcbc9ffddb3.png", "website": "https://kebra.com", "all_locations": "San Francisco, CA, USA", "long_description": "Kebra builds AI for skilled technicians.\r\n\r\nKebra guides technicians in the field in real time, documents their job, and automatically handles hours of back-office work; from dispatching technicians to filing warranty claims and managing inventory.", "one_liner": "AI For The Skilled Workforce", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Operations", "launched_at": 1784399832, "tags": [ "B2B", "Home Services", "Operations", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Operations" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/kebra", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/kebra.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/kebra-com/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/2e21f70e5b9032365254edae847e8dcbc9ffddb3.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1693074, "full_name": "Luca Hadife", "title": "Founder/CEO", "founder_bio": "Co-Founder and CEO of Kebra. Built multimodal AI systems at Microsoft. Early team member at Amata (Series A). UC Berkeley Data/ML and ASUC Senator", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/luca-hadife/", "x_url": "https://x.com/LucaHadife", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ce9563442481e75ad730242772f72ce549481e01.jpg" }, { "user_id": 2600431, "full_name": "Mohamed Hosny", "title": "Founder/CTO", "founder_bio": "Co-Founder and CTO at Kebra. Taught Computer Architecture and Signal Processing at UC Berkeley EECS. Software Engineer on data center switches at Arista Networks.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/mohamed-hosny-hussein/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1a8e7ab2dd7a5e9830daa8cf984167f1a15a83a2.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108355, "title": "Kebra — Making the field queryable", "tagline": "We capture what happens in field operations and turn it into the context AI employees need to automate the work that follows", "body": "Hey YC! 👋\n\nWe’re [Luca](https://www.linkedin.com/in/luca-hadife/) and [Hosny](https://www.linkedin.com/in/mohamed-hosny-hussein/), founders of [Kebra](https://kebra.com/?utm_source=ycombinator&utm_medium=launch&utm_campaign=yc_launch_2026).\n\n## TL;DR\n\nKebra makes the field queryable.\n\nWe capture what happens across jobs, sites, and assets and turn it into a live operating memory that AI can query and act on. Our AI employees use that context to automate the work that follows across the systems companies already use.\n\nEvery job makes that memory richer, building the company specific context software agents use today and physical agents will need tomorrow.\n\nWe’re already working with field operations companies across Florida, Pennsylvania, and California.\n\n\n\n## The problem\n\nField operations never got their Claude Code moment.\n\nWe spent months inside HVAC companies trying to automate their back office. We built AI employees for claims, reporting, coordination, and follow ups.\n\nThey kept breaking.\n\nNot because the AI could not do the work, but because it was acting on a reality that never happened. The most important part of the operation happened on the ground and was never properly captured.\n\nA missing serial number. A photo never taken. What the technician actually found stuck in their head.\n\nYou cannot automate an alternate reality. Before AI can operate the field, the field has to become queryable.\n\n## Our Solution\n\nKebra makes field operations queryable.\n\n**Capture what happens:** Field teams document jobs through photos, voice, video, and structured workflows. Kebra checks what is missing while they are still on site.\n\n**Build the operating memory:** Every job updates a live history of sites, assets, repairs, decisions, and how the company actually works.\n\n**AI employees act on it:** They use that context to automate reports, claims, coordination, compliance, follow ups, and other back office work across the systems the company already uses.\n\n## Vision\n\nEvery job makes Kebra’s operating memory richer.\n\nToday, software agents use that context to automate field operations. Tomorrow, physical AI and robots will need the same company specific context to understand each site, asset, procedure, and history before they can work autonomously.\n\nKebra is building the context layer for field operations.\n\n## Our Ask\n\nRun or know a company with field operations where important information gets lost between the field and the office?\n\nEmail us at [founders@kebra.com](mailto:founders@kebra.com) and [book a demo](https://calendar.app.google/s1DwDxHTbxZ9w5Uq6).\\\nFollow us on [LinkedIn](https://www.linkedin.com/company/kebra-com)", "url": "https://www.ycombinator.com/launches/SBf-kebra-making-the-field-queryable", "votes": 13, "created_at": "2026-07-30T22:13:43.328Z" } ] }, { "id": 31858, "name": "Torus", "slug": "torus", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/79a2f7b5a083998db3fcc9309c0435a2ccb05016.png", "website": "https://www.usetorus.com", "all_locations": "San Francisco, CA, USA", "long_description": "Torus lets engineering firms and project owners build infrastructure 10x faster to reindustrialize America and the West. \r\n\r\nWe're working with multiple Fortune 500 companies to build critical infrastructure including data centers and critical energy installations. ", "one_liner": "Legora for physical engineering firms", "team_size": 6, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1783552199, "tags": [ "Construction", "Energy", "Enterprise Software", "AI", "Industrial" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/torus", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/torus.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/torus-intelligence/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/79a2f7b5a083998db3fcc9309c0435a2ccb05016.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 554410, "full_name": "Marcus Lima", "title": "Founder/CEO", "founder_bio": "Co-founder and CEO of Torus. We build agentic engineering tools for capital project engineering firms.\n\nPreviously founder of Heimdal, direct air carbon capture technology and infrastructure.\n\nOxford and Cambridge grad.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/marcus-lima-144974101/", "x_url": "https://twitter.com/MarcusLimaGN", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/48308bbcb3e74764b7772075c1ed7357f112d2d7.jpg" }, { "user_id": 1208048, "full_name": "Rahul Thayil", "title": "Founder/CTO", "founder_bio": "Co-founder & CTO at Torus (YC S26). \n\nPrev. Founder of Orion - AI threat intelligence and response automation.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rahul-thayil/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d691d142de70dcee7c6c00f5b37675b9753999e4.jpg" } ], "jobs": [ { "id": 105565, "title": "AI Engineer (interested in physical infrastructure)", "url": "https://www.ycombinator.com/companies/torus/jobs/jIERZ3O-ai-engineer-interested-in-physical-infrastructure", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$150K - $250K", "equity_range": "0.25% - 0.50%", "experience": "1+ years", "visa": "US citizen/visa only", "skills": [ "Python", "Rust", "TypeScript", "Torch/PyTorch", "Machine Learning", "CAD Design", "LLMs", "RAG", "AI Agents" ] }, { "id": 105567, "title": "Electrical Engineer ", "url": "https://www.ycombinator.com/companies/torus/jobs/kNWpYEu-electrical-engineer", "location": "San Francisco, CA, US / Houston, TX, US", "type": "Full-time", "role": "Engineering", "role_type": "Electrical", "salary_range": "$100K - $200K", "equity_range": "0.25% - 0.50%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [] }, { "id": 105568, "title": "Physical Infrastructure Engineer (any discipline)", "url": "https://www.ycombinator.com/companies/torus/jobs/7TMYD7B-physical-infrastructure-engineer-any-discipline", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Chemical engineering", "salary_range": "$100K - $200K", "equity_range": "0.25% - 0.50%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [ "CAD Design", "AI Agents" ] }, { "id": 105570, "title": "Member of Nontechnical Staff (GTM)", "url": "https://www.ycombinator.com/companies/torus/jobs/kBIlCwY-member-of-nontechnical-staff-gtm", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Marketing", "role_type": null, "salary_range": "$100K - $150K", "equity_range": "0.05% - 0.10%", "experience": "1+ years", "visa": "US citizen/visa only", "skills": [ "Marketing design" ] } ], "news": [], "launches": [ { "id": 114335, "title": "Torus: Legora for the engineers who build power plants, refineries, and data centers", "tagline": "better engineering design for industrial infrastructure", "body": "**TL;DR** Torus is Legora for physical engineering firms\n\nWe make engineering design better. Our customers are using Torus to generate complete engineering packages and artifacts including complex engineering diagrams across disciplines. This dramatically cuts engineering cost, and timelines. \n\n![uploaded image](/media/?type=post&id=114335&key=user_uploads/554410/77f6d39c-d4dd-40d5-9c3d-af74b6daa8d9)\n\nInfrastructure projects spend months or years just on engineering. Every week of schedule compression on billion dollar infrastructure projects is worth millions. \n\nWe’re working with major EPCs and infrastructure owners across the US, helping them cut months long engineering timelines down to days. AI is fundamentally transforming how physical engineering is done, and Torus is delivering this to customers. We’re doing to physical enmgineering what Cursor did to software engineering. \n\n![uploaded image](/media/?type=post&id=114335&key=user_uploads/554410/0118fdd5-ef91-4497-9f28-858e4a6c0b77)\n\nMarcus Lima, CEO. Founded Heimdal (YC S21), raised $12M+ including Sam Altman and Marc Benioff, and built the world's largest carbon capture plant of its kind, working directly with EPCs for years. He lived the pain of slow and expensive engineering.\n\nRahul Thayil, CTO. Founded Orion, raised $4M+, and spent 10+ years building real-time anomaly detection over messy physical data.\n\nBest friends for 13 years, lived together for 5. One of us built the infrastructure; the other built the intelligence over its data. Torus is both, fused.\n\n**Our ask**\n\n* **We're hiring founding engineers (SF, in person).** We're based in SF and work in person, and we forward-deploy to our customers in Kansas City, Tulsa, Houston, and wherever the projects are. We’re hiring every type of engineer, software, mechanical, electrical, process. If you’re an engineer interested in building massive infrastructure we want to work with you.\n* **Intros to infra owners / engineering firms** the companies that build out infrastructure (midstream, power, utilities, data center developers, industrial operators, O&G). Anyone scaling capacity who feels the engineering crunch.", "url": "https://www.ycombinator.com/launches/Tk7-torus-legora-for-the-engineers-who-build-power-plants-refineries-and-data-centers", "votes": 6, "created_at": "2026-09-09T20:00:00.280Z" } ] }, { "id": 31923, "name": "tash", "slug": "tash", "former_names": [ "tash." ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/8fc74afee1c19a4cdde9eb90d55141af2d3baba7.png", "website": "https://www.tradetash.com", "all_locations": "San Francisco, CA, USA", "long_description": "tash is building the investment platform for sports and trading cards. We create professionally constructed portfolios of iconic, investment-grade cards, giving investors exposure to the asset class without the hassle of selecting, buying, storing, and selling individual cards.\r\n\r\nTrading card investing remains fragmented, opaque, and difficult to access. tash is building the infrastructure to make the asset class simpler, more transparent, and professionally managed. We are currently pursuing SEC qualification to launch regulated investment offerings backed by trading cards.", "one_liner": "The S&P of Trading Cards", "team_size": 4, "industry": "Consumer", "subindustry": "Consumer", "launched_at": 1785522137, "tags": [ "Consumer", "Investing" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Consumer" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/tash", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/tash.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/tash-cards/", "x_url": "https://x.com/tashcards?s=11", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/8fc74afee1c19a4cdde9eb90d55141af2d3baba7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3101812, "full_name": "John Ohannesian", "title": "Founder/CEO", "founder_bio": "Co-Founder @ tash, Ex-IB & UC Berkeley, working towards bringing collectibles to the public markets ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/johnohannesian/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c6b96e96f5e445dda3caf1f2abff7f59f82c9287.jpg" }, { "user_id": 1806514, "full_name": "Nathan Borshell", "title": "Founder/CTO", "founder_bio": "Co-Founder at tash (S26). Previously at UBS Global Markets and Seso. Economics and Data Science at UC Berkeley.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/nathanborshell/", "x_url": "https://x.com/nathanborshell", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3b1c74e8d3e948ae7ca10f901214f64bf8331170.jpg" }, { "user_id": 3232492, "full_name": "Victor Shammas", "title": "Founder/CPO", "founder_bio": "UC Berkeley Business & Data Science, Co-Founder @ tash, building the public markets for collectibles.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/victorshammas/", "x_url": "https://x.com/victorshammas", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b7cbabaeccca0ca1b75e8a04df418281e21aece7.jpg" }, { "user_id": 3063171, "full_name": "Nicolas Sharma", "title": "Founder/COO", "founder_bio": "Co-Founder @ tash. Previously at J.P. Morgan and Capgemini. Econ + Business + Data Science at UC Berkeley.\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/nicolas-sharma/", "x_url": "https://x.com/nico_sharma", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2e299adf7def9e2b63edc2ed408d2da52f937ade.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109453, "title": "tash - Making the world’s most iconic trading cards investable", "tagline": "Building the financial infrastructure to bring collectibles into public markets. Starting with trading cards.", "body": "**A $50B+ asset class is still being invested in one card at a time. We’re building the investment platform for trading cards.**\n\n**TL;DR:** tash gives investors access to professionally curated portfolios of high-end graded trading cards. tash manages sourcing, verification, vaulting, reporting, and the eventual sale of the underlying assets.\n\n[https://drive.google.com/file/d/1bszOZp6uTWIgG29grz1oxflgInfEvpD6](https://drive.google.com/file/d/1bszOZp6uTWIgG29grz1oxflgInfEvpD6/view)\n\n## The problem\n\nTrading cards have evolved into a global asset class supported by professional grading, extensive sales data, insured vaulting, and multimillion-dollar transactions.\n\nBut investing has barely changed.\n\nToday, investors still have to research, source, price, verify, store, track, and eventually resell every card individually. Building diversified exposure often means becoming a part-time card dealer.\n\nReal estate has REITs. Gold has ETFs. Equities have index funds.\n\nThe asset class exists. The investment product does not.\n\n## What tash does\n\ntash builds professionally managed trading card portfolios around distinct parts of the market.\n\nInitial portfolios range from broad market exposure through the tash25 Index to focused strategies like GOATs, Rising Stars, Vintage, Pokémon, and TCGs.\n\nInvestors gain diversified exposure while tash manages sourcing, verification, secure vaulting, insurance, reporting, and the eventual sale of the underlying cards.\n\ntash is actively pursuing SEC qualification to make these products available to more investors.\n\n## Why us\n\nWe grew up collecting cards and never stopped.\n\nVictor, Nathan, and John met in high school before attending UC Berkeley, where they met Nico and became roommates. Together, we have bought and sold more than $1M of trading cards and collectibles.\n\nOur professional experience spans finance, data, and marketplaces, alongside years spent buying at card shows, negotiating, tracking prices, and studying the market.\n\n## The broader vision\n\nTrading cards are the starting point.\n\ntash is building the financial infrastructure needed to bring the broader $330B+ collectibles market into public markets, making cultural assets easier to access, evaluate, and invest in.\n\nInterested in investing in trading cards? Reach out to the team or join the waitlist at [**tash.cards**](http://tash.cards).\n\nContact: [founders@tash.cards](mailto:founders@tash.cards)\n\n_For discussion purposes only. This is not an offer to sell or a solicitation to purchase securities. Any offering will be made solely through definitive offering documents and in accordance with applicable securities laws. Investments involve risk, including possible loss of principal._", "url": "https://www.ycombinator.com/launches/STN-tash-making-the-world-s-most-iconic-trading-cards-investable", "votes": 7, "created_at": "2026-08-06T20:32:51.338Z" } ] }, { "id": 31989, "name": "Pango", "slug": "pango", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/d6fb7711e26bf57516e7dcfa35cfadeb5d98b5b2.png", "website": "https://pango.ai/", "all_locations": "New York City, NY, USA", "long_description": "Pango is a platform for e-commerce companies to use natural language to build agents and automate their logistics operations like delivery and returns.", "one_liner": "Agentic OS for e-commerce operations", "team_size": 5, "industry": "B2B", "subindustry": "B2B -> Retail", "launched_at": 1781764394, "tags": [ "Logistics", "E-commerce", "Retail", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Retail" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/pango", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/pango.json", "year_founded": 2025, "city": "New York City", "country": "US", "linkedin_url": "https://www.linkedin.com/company/pangoai", "x_url": "https://x.com/Pango_ai", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/pango-ai", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/d6fb7711e26bf57516e7dcfa35cfadeb5d98b5b2.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 3288698, "full_name": "Steve Rahimi", "title": "Founder/CEO", "founder_bio": "Founder at Pango (YC S26). E-commmerce entrepreneur since 17. 7 years of building and scaling in e- commerce and tech.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/steve-rahimi/", "x_url": "https://x.com/Steve_rahimi", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/171d14e5486ca0d0a6ec63893fda171ebe2820de.jpg" }, { "user_id": 3331619, "full_name": "Lukasz Reszczynski", "title": "Founder/CTO", "founder_bio": "Founder at Pango (YC S26). Previously Lead Developer at Partfiniti where he created a patented configurator that powered over $1.5B in product quotes and the entire sales process of Accu-Tech ($300M revenue). Acting CTO for several companies, led delivery infrastructure development for a UK restaurant chain. MSc in Astrophysics from Liverpool John Moores University, from which he pivoted to PhD in Molecular Biology from Universität Wien, before dropping out to start Pango. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/lukasz-reszczynski-bio/", "x_url": "https://x.com/_lukres", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9f1ff89daf4143621266c9f4ab07922a3e22787e.jpg" } ], "jobs": [ { "id": 112412, "title": "Founding Engineer", "url": "https://www.ycombinator.com/companies/pango/jobs/xXKwYhq-founding-engineer", "location": "Stockholm, Stockholm County, SE", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$36.2K - $145K", "equity_range": "0.20% - 2.00%", "experience": "3+ years", "visa": "Will sponsor", "skills": [] } ], "news": [], "launches": [ { "id": 107204, "title": "Pango: E-commerce operations on autopilot.", "tagline": "Why can't e-commerce companies scale like tech companies?", "body": "tl;dr: Why can't e-commerce companies scale like tech companies? And how we might be able to change that.\n\nWe spent 12 months embedded in the operations of 30 e-commerce brands\\! Here’s what we learnt:\n\n![uploaded image](/media/?type=post&id=107204&key=user_uploads/3288698/1a3cc8c3-db71-4d81-b2ad-9679b77f9451)\n\n**What we saw:** Each company had a unique operational setup with a mix of 5 to 7 tools glued together to run the shop. To keep all these tools running smoothly with each other, it requires a lot of tedious admin. As a result, headcount increases as the company scales.\n\nMost companies couldn't even automate a basic return processing, let alone offer exchanging items or build an AI agent to approve returns or send shipping label to the customer.\n\nExpansion made it worse: more countries and/or SKUs = exponentially more problems.\n\n**What we built:** We have built an agentic OS for e-commerce: You connect your shop stack, then describe how you want to handle shipments, returns, refunds, or claims → Pango builds the return portal, fixes the delivery options, and builds a customer service process and agent for you.\n\n![uploaded image](/media/?type=post&id=107204&key=user_uploads/3288698/ceed8a01-52aa-403a-9986-758da64ff4e4)\n\n**Result:** So far we have managed to automate 99% of returns and shipping operations, reduce operation costs by 20% and sign 30 global brands.\n\nAsk\n\nIntros to:\n\n- Agencies (Shopify/e-commerce)\n- Brands in fashion, beauty, home.\n- Forwardable: \"Pango runs everything after checkout: tracking, returns, and exchanges in one platform. Demo: \" Or book us directly: Or just email me if you are intrested: \n", "url": "https://www.ycombinator.com/launches/Rt6-pango-e-commerce-operations-on-autopilot", "votes": 10, "created_at": "2026-07-24T14:19:26.215Z" } ] }, { "id": 32089, "name": "Hoplite", "slug": "hoplite", "former_names": [ "Carbon Copy", "Carboncopy", "Blueprints" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/087c3f8fd341b06cc103d1bfc1f15c78f2464f4a.png", "website": "https://hoplite.sh", "all_locations": "San Francisco, CA, USA", "long_description": "", "one_liner": "Effortlessly deploy cloud software factories.", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1782199103, "tags": [ "Developer Tools", "Cloud Computing", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/hoplite", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/hoplite.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/hoplite-sh", "x_url": "https://x.com/hoplite_sh", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/087c3f8fd341b06cc103d1bfc1f15c78f2464f4a.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 2563241, "full_name": "Ryan Morrissey", "title": "Founder/CEO", "founder_bio": "Co-Founder & CEO at Hoplite, Prev. Stripe, Induct ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ryan-morrissey-834256271/", "x_url": "https://x.com/ryan_morr", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/70a86a38d371096d8819303d0cf325e7b22af67d.jpg" }, { "user_id": 2614746, "full_name": "Bence Redmond", "title": "Founder/CTO", "founder_bio": "Founder & CTO at Hoplite, prev. AWS", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/bence-redmond", "x_url": "https://x.com/benceredmond", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c2283c4810bf71aac47f96aad681ac05c0a0eb68.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108274, "title": "Hoplite - Effortless cloud agents for programming", "tagline": "Hoplite runs autonomous coding agents in cloud sandboxes. Spin up tasks manually from Slack, Linear, iMessage, or automatically from Sentry.", "body": "Cloud coding agents are painful to set up, lacking ergonomics, and don't allow you to move as fast as possible. We let you bring your local set up to the cloud.\n\n![uploaded image](/media/?type=post&id=108274&key=user_uploads/2563241/28db62de-d65b-4498-84ac-12e2683c65f2)\n\n**The Problem** \n\nIn the next 10 years the best software startups will be those who can effectively deploy tokens. \n\nRight now that's significantly harder than it needs to be. If you’ve ever ran multiple agents locally you immediately know this pain.\n\nLaptop crashes, wrangling gitwork trees, juggling ports, and more headaches.\n\n![uploaded image](/media/?type=post&id=108274&key=user_uploads/2563241/6f320e85-ccbb-4fb4-a5e9-6ec3cb0206c7)\n\nSo cloud is the obvious answer right? \n\nBut setting up with these existing providers is quite a hassle, and oftentimes they miss some key feature that your local config has.\n\n**What we do**\n\nWe deploy software factories into businesses that want to **tokenmaxx** but don't have the time or the resources to set up the infrastructure.\n\nDuring onboarding, we transfer all of your local sessions and MCP servers to Hoplite, meaning you can pick up right from where you left off. We also configure all the required dependencies and CLIs that your projects require, letting you see instant previews of what your agent is doing.\n\nOur bet is that as models improve, we’re going to start caring less about the code that they output (we’ll implicitly trust it), and shift our focus towards the product itself. Hence, a platform that prioritizes iteration on product rather than code will ultimately win. \n\nWe also have an iMessage integration so you can prompt on the go.\n\nWe installed in our first business last week and we saw an immediate result, from sentry logs going from largely ignored to fully proactive PRs when something breaks. \n\nThey have significantly faster iteration cycles and lower priority tickets have been continuously getting merged. \n\nThey wanted faster iterations but didn't have the infrastructure to support a software factory, so we went in and built it.\n\nIf this sounds like you reach out! \n\nP.S. – If you love Vim, this is for you. It was built with keyboard navigation in mind 🙂\n\ndemo vid: \n\nsign up: \n\n**How we got here**\n\nPreviously, we were building an agentic harness for financial markets; letting retail investors trade with AI. The product worked well, but we never truly had founder market fit. \n\nWe found that our true passion was in talking to founders and developers – and oftentimes when pitching our old product, I’d spend 5x longer explaining our potential pivot 😅\n\nAnd so we set out to commercialize a product that we’d built internally, one that’d let us deploy the $1.5 million in OpenAI credits effectively.\n\n**Our Ask:** Let us Collison install! Bence and I will come to your apartment and set up everything for you (and bring some matcha & assorted sweet treats). We want to build the best possible product for developers to move faster and care deeply about peoples setups. Partially because we have very strong opinions of our own… \n\n**The Team** \n\nBence & I are both from Ireland 🇮🇪. We met in college in a course called Immersive Software Engineering. \n\nWe have been building projects & doing freelance work tg for the past 2 years, but got sick of building other people's visions. \n\nWe previously interned at AWS (Bence) & [@Stripe](https://bookface.ycombinator.com/company/240) (Ryan). After that experience, we were fed up and decided to drop out to build. \n\nWe are best friends & just want to build a product that people truly love. ", "url": "https://www.ycombinator.com/launches/SAM-hoplite-effortless-cloud-agents-for-programming", "votes": 11, "created_at": "2026-07-30T18:32:37.080Z" } ] }, { "id": 32121, "name": "Moonshot", "slug": "moonshot-computer", "former_names": [ "Moonshot", "OpenTrade" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/609d8457269ad1be9496a28a67f4cd8fd8c26af1.png", "website": "https://moonshot.computer/", "all_locations": "San Francisco, CA, USA", "long_description": "We want to bring AI agents to a billion people.\r\n\r\nTo do this, we believe it needs to be 1. proactive and 2. personal. A billion people already carry the right device, but none of them have the right connector.\r\n\r\nMoonshot is personal intelligence that lives on your iPhone and listens to you 24/7. It continuously builds up a greater understanding of you through constantly listening and building context. It just pays deep attention, the way a good friend does. Through this, Moonshot monitors your conscious and subconscious life, figuring out what you need before you know, reminding you about open loops, identifying life and emotional trends before you recognize it, and covering a lot more of what we gloss over in daily life.\r\n\r\nFor fifty years, software has waited to be asked. We think tools are just good enough to flip this, and the next great software will provide value before asking for something first. We're doing it through Moonshot.\r\n\r\nOur thesis is simple: you're only living half your life. The half that you actively remember. The other half just evaporates into your subconscious: what you said, felt, promised, and forgot. It's the more interesting half, and until now nothing could catch it, making it not actionable.\r\n\r\nMoonshot catches your subconscious half and understands you well enough to take actions on your behalf.\r\n\r\nTake life personally.\r\n\r\nThe iOS app is designed around ambient listening and clear user control. Short microphone chunks are sent securely to Moonshot's private transcription service and deleted when transcription finishes. Text and selected context can be sent through Moonshot's server to a model when the app needs to understand a moment. Nothing leaves without your yes. The privacy policy explains every data path, stored copy, permission, and deletion control in detail.\r\n\r\nMoonshot is in early access. Join the waitlist to hear when access opens, read more about the team and product on the About page, or use the Contact page for a direct written response from the Moonshot team.\r\n\r\nLet's make it a Moonshot.", "one_liner": "agents for the rest of us", "team_size": 3, "industry": "Consumer", "subindustry": "Consumer", "launched_at": 1783572660, "tags": [ "Consumer", "Productivity", "AI", "AI Assistant" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Consumer" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/moonshot-computer", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/moonshot-computer.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/heymoonshot", "x_url": "https://x.com/heymoonshot", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/609d8457269ad1be9496a28a67f4cd8fd8c26af1.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 2384611, "full_name": "Ojas Kandhare", "title": "Founder/CEO", "founder_bio": "goofing", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ojas-kandhare-74b349212/", "x_url": "https://x.com/ojaskandy", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5ce3ae32b375e26e931c634c3b08c5c3e3281e32.jpg" }, { "user_id": 2613840, "full_name": "Tario You", "title": "Founder/CTO", "founder_bio": "mooning", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/tario-you", "x_url": "https://x.com/deftarioo", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8de95ae9d0f4ced5184d6984735de601565c588c.jpg" } ], "jobs": [], "news": [], "launches": [] }, { "id": 32122, "name": "Donkey", "slug": "donkey", "former_names": [ "Prelude" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/8bacc206e6479ad9ca5809a5242ecdf78c6f5535.png", "website": "https://donkey.trade", "all_locations": "San Francisco, CA, USA", "long_description": "Donkey is the AI-native trading company. Upload a buying list and we land the whole container at one locked, delivered price, tariffs included. You buy from Donkey: one contract, one company responsible for everything between the factory floor and your door. Our own people check the cartons before the factory gets paid, and the balance is due when the goods are accepted at your dock.\r\n\r\nSmall businesses buy imports through three or four middlemen, each adding 15 to 30 points, because none of them can fill a container on their own. Big enough to buy pallets, never big enough to buy the box. Quotes float with freight, FX and tariffs and live 15 to 30 days.\r\n\r\nA trading company is rooms of people on the phone with factories, forwarders and brokers. Donkey runs that as one engine. It finds the real factory behind the customs record, stacks duty line by line, packs the container on weight and cube, and returns one number that holds from signature. Next comes the shared container: small buyers fill one box together at container prices, and every container earns better factory terms for the next one.\r\n\r\nTeam in San Francisco and Wuhan, English and Mandarin. If your business buys pallets of anything imported, send us your buying list.", "one_liner": "The AI-native trading company: factory prices, delivered, one number", "team_size": 5, "industry": "B2B", "subindustry": "B2B -> Supply Chain and Logistics", "launched_at": 1784315538, "tags": [ "Logistics", "Manufacturing", "Supply Chain", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Supply Chain and Logistics" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/donkey", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/donkey.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/thedonkeytrade/?viewAsMember=true", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/8bacc206e6479ad9ca5809a5242ecdf78c6f5535.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/83521fec3b262b8bc72bd6db93e75881b4fb1a0a.png" ], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 3038319, "full_name": "Benjamin Martindale", "title": "Founder/CEO", "founder_bio": "Grew up in his family's importing business in Toronto. Advised Chinese heavy-machinery giants on selling into the West and ran a Wuhan office in Mandarin. Featured on Mango TV, one of China's largest broadcasters: factory-floor segments and a named 13-minute profile. Founded a Toronto marketing company: $4M of construction deals closed in six months. Finance at Rotman, UofT, one credit short.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/benjamin-martindale/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/045809463bf1d66616681e28482ee5bad8d1f20a.jpg" }, { "user_id": 3041470, "full_name": "Minghao Tan", "title": "Founder/CTO", "founder_bio": "Co-founder & CTO. Donkey buys factory-direct in China and sells delivered to US importers. Customs records are public. We read them, flag the trading company markup, and quote the same parts straight from the factory that makes them. QC on the floor before the factory gets paid.\nI build the data side: parsing raw customs filings, separating real factories from shells and forwarders, and the software the whole trade runs on. \ndonkey.trade", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/james-tan-5a4b07224/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/81a0230e7bc211c76037d8cb1aeb802632abf5c4.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111184, "title": "Donkey — the AI-native trading company", "tagline": "The AI-native trading company: factory prices, delivered, one number.", "body": "Everything gets cheaper by the container. Big companies buy the box. Small businesses buy from the people who bought the box.\n\nDonkey turns the container into a cart. Upload your buying list and see every landed price instantly, tariffs and delivery included. Our engine packs the order, and when the load needs a bigger box, a second container steps in. When you sign, the price holds.\n\nYou buy from Donkey. One contract, one company responsible for everything between the factory floor and your door. If costs move under your locked price, that's our problem. And our own people are on those factory floors, checking what goes into the box before it ships.\n\nOur AI agents read public customs records to find the real factories and pull their quotes. Duty gets stacked line by line. The pricing math is deterministic, not AI, so the number you see is the number you pay. Factory pricing and terms improve with volume, and Donkey's volume grows with every container.\n\nA home builder in Oakland put 15 SKUs into one container with us and saved 38% versus buying piecemeal from distributors. We're starting with the people who build.\n\nWe're Ben and James. Ben grew up in his family's industrial supply business and helped Chinese industrial giants like Sany and Zoomlion sell into Western markets. James built agent systems for industrial supply chains at InstaLILY. We’ve lived both ends of the container.\n\nIf your business buys inventory, bring us your buying list. [donkey.trade](http://donkey.trade)\\\n", "url": "https://www.ycombinator.com/launches/SvI-donkey-the-ai-native-trading-company", "votes": 9, "created_at": "2026-08-19T18:00:00.190Z" } ] }, { "id": 32123, "name": "HERA", "slug": "manufacturingintelligence", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/caa902826145e78311e27007937976e8663a9265.png", "website": "https://www.manufacturingintelligence.org", "all_locations": "San Francisco, CA, USA", "long_description": "At Hera, we build tools that help heavy engineering teams bring products to market up to 20× faster. Today, we work with manufacturers of data center ancillary equipment, renewable energy systems, robotics, heat exchangers, power plants and pressure vessels.\r\n", "one_liner": "We make sure every factory in the world builds without mistakes.", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1784416117, "tags": [ "B2B", "Design", "Design Tools", "Manufacturing", "Industrial" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": [ { "question": "What is the core problem you are solving? Why is this a big problem? What made you decide to work on it?", "answer": "American manufacturers lose an estimated $285,000 annually to poor quality for every $10 million in revenue. At the same time, the industry is losing the engineers who know how to prevent those losses. Nearly 4 million manufacturing jobs will need to be filled by 2033 as experienced workers retire, and almost half are expected to go unfilled. Much of the expertise required to review complex engineering drawings exists only in the heads of senior engineers who are leaving the workforce.\r\n\r\nWe're both engineers who grew up watching this firsthand. After more than 200 conversations with manufacturing engineers, quality engineers and machinists, we're building software that captures that expertise by automatically reviewing engineering drawings against ASME standards and company-specific rules before they ever reach the shop floor." }, { "question": "What is your long-term vision? If you truly succeed, what will be different about the world?", "answer": "We're building the largest repository of engineering judgment in the world. We tailor each company's checks to their own standards, the kind usually mastered by only a handful of senior engineers. We turn that knowledge into software every plant can use the same week.\r\n\r\nWith success, the physical things the world depends on get built 10x faster, and a retiring generation's expertise is never lost." } ], "url": "https://www.ycombinator.com/companies/manufacturingintelligence", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/manufacturingintelligence.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/heraengineer/posts/?feedView=all", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/caa902826145e78311e27007937976e8663a9265.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/0f48ca5529aa1b18d9d6d0b656ed5b0bcfae271f.png", "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/dda0e108df1d005a8106e912cf690d30e4643294.png" ], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 1972660, "full_name": "Meera Patel", "title": "Founder/CEO", "founder_bio": "CEO of HERA. Previously delivered projects for Pfizer and TE Connectivity, and built factory automation systems used by millions. ML research and Industrial Engineering at UW.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/meera-patel-92704b207/", "x_url": "https://x.com/lockedincheetah?s=21", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4f51e71bee8d216d4bfee5f0f7eb14687afb9605.jpg" }, { "user_id": 3352287, "full_name": "Noelle So", "title": "Founder/CTO", "founder_bio": "Co-founder of HERA. Built water infrastructure at 19 serving companies behind 6% of Philippine GDP. Previously developed software for engineering facilities at TE Connectivity & Wyze Labs. Former machine learning and industrial engineering researcher at UW.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/noelle-so-055474140/", "x_url": "https://x.com/noelleyso", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2a0cf52a331c47b6a753767b0bdf41c846502342.jpg" } ], "jobs": [ { "id": 108320, "title": "Industrial Software Engineer", "url": "https://www.ycombinator.com/companies/manufacturingintelligence/jobs/ffZp9fi-industrial-software-engineer", "location": "San Francisco, CA, US / Remote (US)", "type": "Full-time", "role": "Engineering", "role_type": "Backend", "salary_range": "$150K - $200K", "equity_range": "", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [ "Amazon Web Services (AWS)", "JavaScript", "CAD Design", "AI Agents" ] } ], "news": [], "launches": [ { "id": 107030, "title": "HERA: catches thousand-dollar drawing mistakes before the shop floor.", "tagline": "Hera reviews every drawing against your factory's own standards and the ASME code, catching expensive mistakes and preventing rework 10x faster than manual review.", "body": "**TL;DR: Every custom metal part starts as a drawing, and every drawing gets checked by hand before anyone builds it. Hera does that check in minutes instead of a full day, and catches the mistakes that turn into scrapped parts.**\n\n[**https://youtu.be/J_5NRjkou3Y**](https://youtu.be/J_5NRjkou3Y)\n\n**The Problem**\n\nCustom mechanical parts run on 2D drawings. Whether you're an OEM releasing drawings to suppliers, or an aerospace supplier answering to source inspection, every drawing has to be checked against ASME/ISO standards and GD&T requirements before it goes anywhere near a machine.\n\nRight now, that review is done by hand, and usually by the most senior engineer in the building. And a single missed tolerance callout or an ambiguous datum reference can mean a scrapped part, a blown schedule, or a shipment that fails inspection.\n\nWe heard the same story from a pressure vessel fabricator, a semiconductor tooling shop, and a medical device manufacturer: reviews take days. This isn't because the work is hard. There's simply no good way to do it faster without cutting corners, and nobody cuts corners on a drawing that could scrap a $40,000 part.\n\n**Our Solution**\n\nHera is a brain for manufacturing. Any real advance in manufacturing has to be grounded in two things: the standards parts are held to and the reality of the shop floor. So that's what we built Hera on.\n\nUpload a drawing and Hera checks GD&T callouts, tolerance stack ups, and title block requirements against ASME Y14.5, relevant ISO standards, and your own design intent. The result is a drawing your shop can build right the first time, without the delays and rework that eat margins. Every review also makes Hera more valuable, creating a system of record for drawing errors that helps teams build better models for engineering processes.\n\n**While the current review takes days, Hera takes minutes… and gets better by the drawing.**\n\n**Us**\n\n**Noelle** grew up in an industrial park her family runs in the Philippines and has been on factory floors her whole life. At 19, she built a water-treatment facility that added 40 L/s of capacity to the park. \n\n**Meera** has spent years around manufacturing: at her family's plants, and on projects at Pfizer and TE Connectivity, where she built a SOTA discrete-event simulation. She taught herself PLC programming and C++, and has published ML research at UW.\n\nWe actually met on a bus to a factory tour, and have been best friends ever since!\n\n![uploaded image](/media/?type=post&id=107030&key=user_uploads/3352287/d05e9e1a-57c8-4e43-91d7-75e42a004261)\n\n**Our Ask** If you run or work at a factory (aerospace, semiconductor, medical device, pressure vessel, or custom machinery), we want to visit you! If you invite us out to see your shop, we’ll come by in person and bring you a **customized Hera Swiss Army knife**.\n\nReach out to us at [**founders@manufacturingintelligence.org**](mailto:founders@manufacturingintelligence.org).", "url": "https://www.ycombinator.com/launches/RqI-hera-catches-thousand-dollar-drawing-mistakes-before-the-shop-floor", "votes": 138, "created_at": "2026-07-24T15:18:23.996Z" } ] }, { "id": 32124, "name": "Erinys", "slug": "erinys", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/3ecec9425d412cfc7fa1460b155883f078a7f023.png", "website": "https://erinys.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Erinys provides the technology, intake, back office, and client acquisition needed to launch and scale an AI-native law firm. Lawyers own their firms and make every legal decision; Erinys handles everything around them.", "one_liner": "We help lawyers start and scale AI-native law firms.", "team_size": 3, "industry": "B2B", "subindustry": "B2B -> Legal", "launched_at": 1786258271, "tags": [ "Legal", "LegalTech", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Legal" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/erinys", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/erinys.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/erinysai/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/3ecec9425d412cfc7fa1460b155883f078a7f023.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3662807, "full_name": "Sasha Zyuzin", "title": "Founder/CEO", "founder_bio": "Founder at Erinys (YC S26). Presented vulnerability discovery research at Black Hat USA 2025 and spoke at the SANS OSINT Summit 2025. Undergraduate researcher at UMD working across IPv6 botnet detection, malware analysis, and hardware security. Previously taught an OSINT course at UMD and led the university's 1800+ member cybersecurity club.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/azyuzin/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f0284a1adccaf56e9d791717ce08f03d2fe4c099.jpg" }, { "user_id": 3680893, "full_name": "Jason Guo", "title": "Founder", "founder_bio": "Founder at Erinys (YC S26). Has experience in very diverse parts of the tech stack, from industrial PLCs to kernel exploitation. In-progress computer engineering degree in the University of Maryland.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jason-guo-225842381/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/796a9fd01db3d5b9c12c815d963a215c4e9e9311.jpg" }, { "user_id": 1511726, "full_name": "Matthew Pallan", "title": "Founder/CTO", "founder_bio": "Founder at Erinys (YC S26). Before Erinys, Matthew led greenfield development at multiple organizations, directing engineering across several product groups. He's a deep technical builder across domains, having engineered sub-290ns trading systems for HFT firms and worked in cybersecurity spanning exploitation and security systems. He scored a perfect 180 on the LSAT and studied Computer Science and Real Estate at the University of Maryland.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/~mp/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/00dcf64db487dfb28e84e5bc0a965dd6082b6756.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109878, "title": "Erinys - Building the first network of AI-native law firms", "tagline": "The infrastructure for exceptional lawyers to build the firms of the future", "body": "**TL;DR:** Erinys is building a network of AI-native law firms. Lawyers bring their expertise, judgment, and client relationships. We provide the technology and business infrastructure that lets them launch a new practice or transform an existing one.\n\n\n\n\\\n**The problem**\\\nLawyers spend surprisingly little of their time doing work that requires their unique expertise.\\\n\\\nRunning a practice means finding clients, managing intake, hiring staff, handling billing, overseeing marketing, and keeping the business moving. Small firms burn lawyer hours on it; large firms hire entire teams to manage this. Either way, it kills time or margins.\\\n\\\nThe legal work itself is still built around lawyers producing everything. Lawyers spend hours turning raw information into research, analysis, and first drafts, even as AI becomes capable of producing much of that work. Existing legal software makes individual tasks faster, but leaves the underlying model unchanged.\\\n\\\nWe think that model should flip. AI should produce the first pass. Lawyers should direct, review, decide, and advocate.\n\n**How we solve it**\\\nErinys provides the infrastructure for lawyers to operate this way from the ground up.\\\n\\\nFor lawyers starting a practice, we help launch it. For lawyers with an existing practice, we help transition it onto the same model.\\\n\\\nOur systems run the business behind the practice while AI assists with fact development, legal research, analysis, and drafting under attorney supervision.\\\n\\\nThe result is a firm where lawyers spend more time exercising judgment and serving clients, and less time producing routine work or managing the day-to-day mechanics.\\\n\\\nOne exceptional lawyer can operate with the capabilities that previously required a much larger organization.\\\n\\\n**Why us**\\\nMost legal AI is being built to make the traditional law firm more efficient. We are building the infrastructure for a different kind of law firm, AI-native at its foundation.\n\nBefore Erinys, our founding team worked in offensive security, another high-stakes, sensitive, and deeply adversarial environment. We are bringing that security-first mindset to an industry entrusted with its clients’ most sensitive information.\n\n**Our Ask**\n\nWe’re looking for exceptional lawyers.\n\nIf you want to start your own practice, or already have one and want to make it AI-native, we want to talk.\n\nEmail us at [founders@erinys.ai](mailto:founders@erinys.ai) or visit [erinys.ai](http://erinys.ai)", "url": "https://www.ycombinator.com/launches/SaE-erinys-building-the-first-network-of-ai-native-law-firms", "votes": 19, "created_at": "2026-08-10T01:00:00.375Z" } ] }, { "id": 32127, "name": "Perceptron ML", "slug": "perceptron-ml", "former_names": [ "Perceptron ML", "Notyfi" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/23f18bbbc3a1b4a54c0c7513b5b49d243b55ff07.png", "website": "https://perceptronml.com", "all_locations": "San Francisco, CA, USA", "long_description": "We build AI systems for law firms, trained on each firm's own matters and deployed privately in their environment: timekeeping, intake and matter monitoring, research, discovery, and drafting. Our grounding engine verifies every fact against a primary source before a model can cite it, which is why firms trust us with work that off-the-shelf legal AI keeps getting wrong.", "one_liner": "We make law firms AI-native.", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1782500527, "tags": [ "B2B", "Legal", "Automation", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/perceptron-ml", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/perceptron-ml.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://linkedin.com/company/perceptron-yc", "x_url": "https://x.com/PerceptronML", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/perceptron-ml", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/23f18bbbc3a1b4a54c0c7513b5b49d243b55ff07.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3087509, "full_name": "Michael Marcotte", "title": "Founder/CEO", "founder_bio": "Founder at Perceptron ML (YC S26). Previously AI Research at NVIDIA, where he built production AI agents for hardware debugging and formal verification. Holds BS in Mathematics and MS in Computer Science from Stanford.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/marcottemichael", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9bf56925080a6f27803cf4df75bcb411a466ad4b.jpg" }, { "user_id": 3122328, "full_name": "Peyton Marcotte", "title": "Founder/CTO", "founder_bio": "Founder at Perceptron ML (YC S26). Previously founded PMARC, developing astronaut exercise equipment for long-duration spaceflight (supported by NASA, Brown University, and 1517 Fund grants). Studying Mechanical Engineering at Brown.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/peytonmarcotte/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/63ff90f1466f6dda0054cdb016ea612d9c73fd6b.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110491, "title": "Perceptron ML: We make law firms AI-native", "tagline": "We help law firms claim the legal work they're leaving on the table by building AI and software around how each firm already works.", "body": "Hi everyone, we're Peyton and Michael Marcotte, brothers and co-founders of [Perceptron ML (YC S26)](https://www.linkedin.com/company/perceptronml/). We make law firms AI-native: we build the AI and software a firm needs to scale. Not a product off the shelf, a system built around how your firm already works, where no fact is used until it traces to a real document.\n\n[Perceptron ML (YC S26) Launch Video](https://drive.google.com/file/d/1c-5RUaBQsJWG-6vJsy0O8PLvAcgpuEco/view)\n\n**The problem.** In 2023, two New York lawyers filed a brief written by AI that cited six cases that did not exist. Courts have caught it more than a thousand times since. So most firms went back to doing everything by hand, and by hand the math is brutal: lawyers lose \\~40% of their day to work that isn't law, and tens of billions in legal work goes unclaimed every year because firms can't move fast enough to take it.\n\n**What we do.** Every firm's bottleneck is different, and it is never the one on the shelf. So we build what each firm actually needs: intake answered and filed at 2 a.m., demands drafted from the record, records requests chased until they land, deadlines and statutes watched, every hour matched to its matter, clients updated by text before they think to call. Every firm gets something different, but underneath it's all the same platform we built once: our grounding engine, our connectors into the firm's systems, and our agent runtime. That's what lets us build this broadly without being an agency.\n\n**Why you can trust it.** The AI we build doesn't guess. Every fact traces to a real document, verified against the primary source before a model can use it. Click anything, and the original opens.\n\n**Why us.** [Michael](https://www.linkedin.com/in/marcottemichael/) did AI research at NVIDIA (hardware-debugging agent in production). [Peyton](https://www.linkedin.com/in/peytonmarcotte/) led two NASA-supported research projects. We come from places where a confident wrong answer is unacceptable, and we started Perceptron ML because law is the biggest industry still run by hand for exactly that reason.\n\n**The ask.** If you know a managing partner, or the attorney who signs your invoices, we'd love an intro: [founders@perceptronml.com](mailto:founders@perceptronml.com)\n\n[perceptronml.com](http://perceptronml.com)", "url": "https://www.ycombinator.com/launches/Sk7-perceptron-ml-we-make-law-firms-ai-native", "votes": 7, "created_at": "2026-08-14T12:00:00.181Z" } ] }, { "id": 32128, "name": "Mass Magnetics", "slug": "mass-magnetics", "former_names": [ "Mass Energy Company" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/26161cef0e683292e39a5da2d6258b38702473a0.png", "website": "https://massmagnetics.com/", "all_locations": "Boston, MA, USA", "long_description": "Mass Magnetics turns recycled rare-earth scrap into high-performance permanent magnets, made in America. The materials we supply are the backbone of electric motors, sensors, speakers, hard drives, and electronics.\r\nOur process enables DFARS-compliant magnets for defense, aerospace, and industrial manufacturers who demand domestic supply without compromising on quality.\r\nOperations in Greater Boston, MA (Greentown Labs) and Columbus, IN.", "one_liner": "USA-made magnetics for robotics and defense. ", "team_size": 3, "industry": "Industrials", "subindustry": "Industrials -> Defense", "launched_at": 1786735839, "tags": [ "Hard Tech", "Robotics", "Drones", "Manufacturing", "Defense" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Defense" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/mass-magnetics", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/mass-magnetics.json", "year_founded": 2026, "city": "Boston", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/26161cef0e683292e39a5da2d6258b38702473a0.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 3652354, "full_name": "Eric Silver", "title": "Founder/CEO", "founder_bio": "Founder at Mass Magnetics. Worked in pilot cathode production for lithium-sulfur battery chemistries. Now I'm fixing American magnet supply chains. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/energyeric/", "x_url": "https://x.com/2998x10tothe8", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c8160b42bdee61a0dff0fef65dd723b8cae326e3.jpg" }, { "user_id": 3637161, "full_name": "Matthew Groll", "title": "Founder/CTO", "founder_bio": "Co-founder and CTO at Mass Magnetics. Former NASA/Columbia/MIT. Previously worked on thermal systems and guidance/navigation at NASA Johnson Space Center for ISS and Artemis lunar missions. Also previously worked as a Member of Technical Staff at Draper Laboratory. Studied Mechanical Engineering at Columbia, captaining the Formula SAE team and Robotics Club, then earned a Masters in Manufacturing from MIT.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/matthew-groll/", "x_url": "https://x.com/matthewxgroll", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/68e632830dda51ab14dec0fac18967c93ed5c8ed.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110611, "title": "Mass Magnetics: USA-made magnets", "tagline": "Vertically integrated magnet manufacturing for aerospace, robotics and defense.", "body": "### **TL:DR; Mass Magnetics makes American magnets in minutes.**\n\n🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲🧲\n\n![uploaded image](/media/?type=post&id=110611&key=user_uploads/3652354/be41c9e5-378d-452f-a111-10380035f14d)\n\n**Every electric motor runs on rare-earth magnets.** EVs, drones, robots, wind turbines, guided munitions: nearly all of them are dependent on Chinese magnets. On January 1st, 20270, [regulation changes will make it illegal to deliver anything that walks, drives, or flies to US Defense if it contains a magnet produced in covered countries like China.](https://www.massmagnetics.com/comply) The domestic supply chain meant to replace it doesn't exist yet. **Mines take a decade to permit, and the ore they pull up leaves radioactive tailings behind, making permitting magnet mining a nightmare.**\n\nWe’re[ ](https://www.linkedin.com/in/energyeric)the founders, [Eric Silver](https://x.com/metalscontent) and [Matthew Grol](https://www.linkedin.com/in/matthew-groll/)l (left and right in the pic below). We’ve known each other since grade school. Since then, Eric’s built multiple battery pilot production lines and materials science labs in the Midwest for NSF-funded companies. Matthew’s built exquisite aerospace gear including Moon-regolith melters and ultra-precise guidance systems for the world’s foremost labs at NASA, Draper, and MIT. **Together we’re making one bet: America's rare-earth problem isn't a mining problem.**\n\n![uploaded image](/media/?type=post&id=110611&key=user_uploads/3652354/2046e9de-2db6-47e8-91fc-651e31e5a834)\n\n**Mass Magnetics’ unfair advantage is sitting in end-of-life electric vehicles, hard drives, and industrial equipment.** The team uses AI models and robotics trained on real scrap to sort metallurgy by exact grade and chemistry. Then, a rapid-sintering process takes that reclaimed mass to a finished magnet in 3 minutes, where conventional approaches take over 4 hours. The result is a magnet as strong as anything imported, with no radioactive footprint or bureaucratic permitting, fully compliant for US defense.\n\n**If you build robots, drones or motors visit our site and reach out at [outreach@massmagnetics.com](mailto:outreach@massmagnetics.com) and get your magnets.**\n\n![uploaded image](/media/?type=post&id=110611&key=user_uploads/3652354/797eff3c-6d9d-4ffc-894b-664aece1ba74)\n\n", "url": "https://www.ycombinator.com/launches/Sm3-mass-magnetics-usa-made-magnets", "votes": 7, "created_at": "2026-08-18T14:30:00.079Z" } ] }, { "id": 32129, "name": "Egoist Machines", "slug": "egoist-machines", "former_names": [ "studysesh" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/cfd986a9205506ed6e96b2903d048ce2aa98626d.png", "website": "https://ego.ist/", "all_locations": "San Francisco, CA, USA", "long_description": "Egoist Machines is building infrastructure for a world where your personal context belongs to you, rather than to the apps and AI companies you use. Today, every assistant and consumer app builds its own version of you. We think that architecture is backwards. Your context should live with you, persist across products, and be available to any model, app or agent you choose.\r\n\r\nAI Passport is the first layer of that infrastructure. It brings the context you generate across AI assistants and everyday apps into one portable profile you control. Tell ChatGPT something and Claude can already know it, for example. Switch to a new product and it can understand you from the first interaction, without the new company needing to own the underlying data.\r\n\r\nApps integrate through Sign in with AI Passport, requesting only the context they need for a specific purpose and duration. Users can see what is being accessed and revoke that access later. Underneath, we have built our own memory infrastructure for fast, compact per-user retrieval and permissioned access.\r\n\r\nWe are distributing AI Passport directly to consumers through first-party products, while integrating it into consumer applications with existing distribution. Every new product that uses Passport gives users another reason to maintain their context there, and every additional Passport holder makes the context layer more useful to developers.\r\n\r\nThe next step is to build a true Egoist Machine: an intelligent system whose memory, incentives and actions are governed by the individual it represents. We call it Digital Domovoy, the hardware extension of AI Passport. Instead of your most personal context ultimately living on somebody else's servers, it lives and is processed locally on hardware you control. AI Passport becomes the permission layer between that private local context and the cloud models and internet services you choose to use.\r\n\r\nOur North Star is the Egoist Machine: an intelligent system whose memory, incentives and actions are governed by the individual it represents.", "one_liner": "Your personal context, securely portable across every AI app.", "team_size": 2, "industry": "Consumer", "subindustry": "Consumer", "launched_at": 1784701469, "tags": [ "Hardware", "Personalization", "Privacy", "Consumer Products", "AI Assistant" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Consumer" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/egoist-machines", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/egoist-machines.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/130104210/", "x_url": "https://x.com/EgoistMachines", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/egoist-machines", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/cfd986a9205506ed6e96b2903d048ce2aa98626d.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 3328552, "full_name": "Erin McGurk", "title": "Founder/CEO", "founder_bio": "Co-founder & CEO of Egoist Machines (YC S26). Grew 0 to 1M+ followers in a year (@erinmerylstudy) and scaled a social enterprise reaching millions. Advised Cambridge and the UK Government on social media strategy. BA (Uni of Cambridge), Land Economy (Economics + Law). Focused on privacy, and social + legal infrastructure.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/erin-mcgurk", "x_url": "https://x.com/erinmerylstudy", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f0cad001b56062b81b2b1049fee039401556e8fc.jpg" }, { "user_id": 1963978, "full_name": "David Khachaturov", "title": "Founder/CTO", "founder_bio": "Founder at Egoist Machines (YC S26). BA/MA (Uni of Cambridge), MSc (Uni of Oxford), PhD (Uni of Cambridge) in Computer Science. Youngest-ever Fellow at St Catharine's College, Cambridge at 25. Previously ML Researcher at OpenAI (PaperBench), Lead AI Security Researcher at FrontierOne. Founded and scaled mazette! (game publisher, 10-title portfolio). Cambridge CS supervisor for 8+ courses, taught 160+ supervision groups and mentored 19 dissertations.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/david-g-khachaturov/", "x_url": "https://x.com/davidobot_", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/414ccf8066649877960bdc6ad659a315ca611119.jpg" } ], "jobs": [], "news": [ { "title": "Business Weekly Awards 2026: 44 shortlisted companies revealed for Cambridge Tech and Life Sciences showcase", "url": "https://www.businessweekly.co.uk/posts/business-weekly-awards-2026-44-shortlisted-companies-revealed-for-cambridge-tech-and-life", "date": "Jul 24, 2026" }, { "title": "Manchester founder moves to Silicon Valley to build AI startup", "url": "https://businesscloud.co.uk/news/manchester-founder-moves-to-silicon-valley-to-build-ai-startup/", "date": "Jul 17, 2026" }, { "title": "Former Altrincham Grammar student secures backing from world's top startup accelerator", "url": "https://altrincham.todaynews.co.uk/schools/2026/07/16/former-altrincham-grammar-student-secures-backing-from-worlds-top-startup-accelerator/", "date": "Jul 16, 2026" }, { "title": "Former Altrincham student builds new AI company in California | Messenger Newspapers", "url": "https://www.messengernewspapers.co.uk/news/26279308.former-altrincham-student-builds-new-ai-company-california/", "date": "Jul 15, 2026" }, { "title": "Y Combinator backs Cambridge-founded Egoist Machines to launch ‘AI Passport’", "url": "https://www.businessweekly.co.uk/posts/y-combinator-backs-cambridge-founded-egoist-machines-to-launch-ai-passport", "date": "Jul 14, 2026" } ], "launches": [ { "id": 111045, "title": "Egoist Machines: user-owned hyper-personalisation infrastructure for Web 4.0", "tagline": "One portable, permissioned memory and context layer for hyper-personalisation across every AI application.", "body": "**TL;DR:** We are building AI Passport, the user-owned context layer for consumer AI. It lets people carry their personal context between products while controlling exactly what each product can access, why, and for how long.\n\nRead our manifesto for the bigger vision: \n\nCreate your AI Passport to start taking control of your data: \n\n**Ask**: if you’re working on a consumer-facing AI application, or know someone who is, we’d love to chat: [founders@ego.ist](mailto:founders@ego.ist)\n\n\n\nThe next era of consumer technology will be **hyper-personalised**. The products that understand each user from the very first interaction will create fundamentally better experiences and, we believe, the next consumer decacorns. People increasingly lack the time and patience to train every new product from scratch.\n\nThe main blocker to ubiquitous consumer AI is that every major AI assistant and consumer app is building its own private and incomplete model of the same person. ChatGPT holds one model, Claude another, while the data in health apps, calendar, shopping history, travel accounts, wearables, and financial tools remain trapped in separate silos. No product has the complete context needed to understand you, and no user has a single place from which to control it.\n\nAs a result:\n\n\\- Users get bored of repeatedly explaining themselves;\n\n\\- Companies spend money building context that already exists;\n\n\\- Personal data remains trapped inside platforms the user does not control.\n\nThe natural end state of this inefficiency is one user-owned memory and context layer that sits above individual products. Apps request the specific context they need, and users grant scoped, time-limited access instead of permanently surrendering their data.\n\n**_That is AI Passport_**.\n\n---\n\n## How AI Passport works\n\nAI Passport is a centralised, user-controlled home for personal context.\n\nYou can think of it as **Link for personal context**. Link lets people securely save their payment and shipping information once and reuse it across businesses; AI Passport does the same for _all_ your context, including preferences, interests, goals, history, health, travel, and so much more.\n\nUsers bring this context into one centralised AI Passport by connecting the apps and AI assistants they already use. They can also save anything they discover online or on social media directly to their Passport. New information enters an approval inbox, where the user decides whether it becomes _private memory_ or part of their _living, shareable public profile_.\n\nThat same Passport then becomes a plug-and-play connector for consumer apps through our **Sign in with AI Passport** OAuth system:\n\n1\\. An app requests the specific context it needs.\n\n2\\. The user sees exactly what is being requested, why it is needed, and how long access will last.\n\n3\\. The user can approve, edit, or deny the request, with the option to grant one-time, time-limited, or continuing access .\n\n4\\. Every permission can later be inspected or revoked.\n\nWe are launching AI Passport directly to consumers and through distribution and deep-integration partners collectively reaching millions of users across dining, ride-hailing, wearables, healthcare, lifestyle, finance and more.\n\nWe built a custom memory backend for AI Passport. It is optimised for compact, exact retrieval across isolated per-user memory stores and supports cryptographic access control at the memory layer. The retrieval path is constrained by the user's active permission, so the product can access only the context authorised for that purpose and period.\n\n---\n\n## Who we are\n\n![uploaded image](/media/?type=post&id=111045&key=user_uploads/3328552/4bb1bba6-8c9f-4d94-b88a-28051f5e6e45)\n\n**Dr David Khachaturov** holds a PhD (Uni of Cambridge) in Computer Science and is the youngest-ever Fellow at St Catharine's College, Cambridge. Previously an ML Researcher at OpenAI (PaperBench) and Lead AI Security Researcher at FrontierOne.\n\n**Erin Meryl McGurk** grew 0 to 1M followers on socials in one year. Advised Uni of Cambridge, UK Govt, OpenAI on social media strategy. Founded social enterprise Girl Economics to 15 countries, 75 ambassadors, 10+ weekly events.\n\n---\n\n## Our Ask\n\n**For users:** create your AI Passport () and tell us what you think!\n\n**For builders:** if you are building a product that could benefit from richer user context, explore our developer docs () and feel free to reach out to [founders@ego.ist](mailto:founders@ego.ist).\n\nWe would especially value introductions to teams building consumer AI, wearables, health products, commerce experiences, or agents that need bounded access to sensitive personal information. If that is you, I want to chat: ", "url": "https://www.ycombinator.com/launches/St3-egoist-machines-user-owned-hyper-personalisation-infrastructure-for-web-4-0", "votes": 59, "created_at": "2026-08-18T14:58:03.221Z" } ] }, { "id": 32130, "name": "Olam Labs", "slug": "olam-labs", "former_names": [ "Gud Labs" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/d6d9543a0527b0b5ac428e043f211c75ac5d1d59.png", "website": "https://olamlabs.ai", "all_locations": "San Francisco, CA, USA", "long_description": "We research complex, high-fidelity simulations for training and evaluating frontier AI models, alongside the infrastructure to test and deploy them at scale. Our research focuses on simulations with realistic socialization, multi-agent interaction, swarms, and ultra-long-horizon tasks.\r\n\r\nTry one of our first releases, Multi-Agent Arena (https://olamarena.com/), where humans come play social strategy games against multiple AI agents. It has multi-agent social games like Poker, Diplomacy, Risk, and more.", "one_liner": "Crafting simulations for frontier lab model training", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1784577901, "tags": [ "Reinforcement Learning", "Gaming", "Data Engineering", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": [ { "question": "How did your company get started? (i.e., How did the founders meet? How did you come up with the idea? How did you decide to be a founder?)", "answer": "Om and Shreshth met in middle school. They were the only two kids in the “gifted” program in their grade, and their first project together was running a snacks and candy store at their school where the profits were used to donate to charity.\\\r\n\\\r\nDuring Christmas holiday in 2025, Om created a multi-agent simulated environment where agents ran their own company in markets competing against each other, and showed it to Shreshth.\\\r\n\\\r\nShreshth, with a background in data and RL among other things, loved the implications of it. Om discussed with him how if multi-agent environments were developed and productionized at scale via frameworks, APIs for infra, and interactive games the kind of data and evaluations it could generate.\\\r\n\\\r\nThe problem was (and still is) that despite the success of multi-agent environments in research, the ecosystem around them has not been pursued to a degree where the utility of multi-agent environments is made readily available to everyone.\\\r\n\\\r\nThey decided to pursue it and applied to YC, and have been in pursuit of a future where we can understand and measure the models that we work with everyday and that will become increasingly autonomous in our everyday lives with AI acceleration. Om and Shreshth both believe there is a severe shortage of good AI evaluations, especially on qualitative behaviors, of which multi-agent environments (and allowing others to build them) is very well suited to solve.\\\r\n\\\r\nMore on Shreshth: [shresh.ca](http://shresh.ca) Shresh’s X: \\\r\nMore on Om: [sensho.xyz](https://www.sensho.xyz/) Om’s X: " } ], "url": "https://www.ycombinator.com/companies/olam-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/olam-labs.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/olam-labs/", "x_url": "https://x.com/olam_labs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/d6d9543a0527b0b5ac428e043f211c75ac5d1d59.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/3e94d77a846523415d54724a99c9965d17599d69.png" ], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 3676793, "full_name": "Om Buddhdev", "title": "Founder/CEO", "founder_bio": "Building simulated worlds for frontier labs to train capable and safe models.\n\nI love learning, tech, competing, and building cool things. I am optimistically AGI pilled and believe most people (even in SF) have not yet internalized what will change over the next 10Y.\n\nPreviously a Staff Engineer @ latitude.io, one of the best Valorant players in the world, math research cited by Terence Tao, TikTok creator /w 190k followers, Roblox game dev, and some other things.\n\nhttps://www.sensho.xyz", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ombud/", "x_url": "https://x.com/sensho", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/91f0defa43559eee579ccd34e4f7fd404cdb2aa6.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109432, "title": "Can you beat frontier LLMs at social strategy games? | Multi-Agent Arena by Olam Labs, evaluating through multi-agent simulations", "tagline": "Play games like Risk, Poker, or Codenames against frontier AIs. We're running public long-horizon games to do evaluations on social behavior and agentic performance.", "body": "**TL;DR:** As many researchers and labs will say, there is a massive shortage of genuinely good AI evaluations. Multi-agent environments are an underdeveloped way of evaluating models that we think will help address our increasingly worse shortage of evals. We're starting with **Multi-Agent Arena**, where anyone can come compete against frontier AI models in social multiplayer games, as we build towards the full stack for multi-agent environments. Try Multi-Agent Arena for free at :) \\\n\\\n[https://drive.google.com/file/d/186GgmnOLz4Yc1BpFk6AxHd3aqUdFYCQy](https://drive.google.com/file/d/186GgmnOLz4Yc1BpFk6AxHd3aqUdFYCQy/view?usp=sharing)\n\n---\n\n## **Overview**\n\nMulti-agent environments provide an essential, underutilized framework for benchmarking autonomous AI models. Multi-Agent Arena is an interactive platform where human participants compete against frontier AI models in multiplayer games, serving as an experimental foundation to evaluate complex agent interactions. Multi-Agent Arena is the first of many stepping stones to building the entire stack and ecosystem for multi-agent environments.\\\n\\\n**The Evaluation Gap** \\\nCurrent AI evaluations primarily focus on single-agent, task-oriented capabilities. While benchmarks such as SWE-bench measure functional task execution, they do not assess the dynamic social behaviors required for autonomous real-world deployment: negotiation, coalition stability, long-horizon planning against adaptive opponents, and spontaneous deception.\n\nSimulated multi-agent environments induce these complex interactions while preserving complete experimental control. Unlike unstructured real-world deployment, these environments offer full data legibility, enabling researchers to inspect every reasoning trace, state transition, and stochastic variable behind model decisions.\n\n**Our Preliminary Evaluations**\\\nThe platform runs deterministic, replayable environments hosting human players alongside anonymized frontier models. Structured as accessible multiplayer games, the infrastructure captures interaction data to score models across two primary axes in our first evaluated game, Poker:\n\n* **Competitive Capability:** One of our preliminary methods is using a Bradley-Terry model to yield Elo ratings with bootstrapped confidence intervals, supplemented by domain-specific metrics such as big blinds per 100 hands in poker. Top-quartile human participants currently outperform all benchmarked models in Poker.\n* **Behavioral Diagnostics:** Model reasoning traces are evaluated against ground-truth game states using expert-curated rubrics and variance-checked grader panels to ensure inter-annotator agreement.\n\n**Initial Metric: Deception Index** \\\nThe benchmark's initial diagnostic metric, the Deception Index, quantifies deliberate misrepresentation per 10,000 turns. A deceptive event is logged only when an agent's private chain-of-thought confirms awareness of a true state while its external communication directly contradicts it. In baseline tests with no explicit instructions or prompts to deceive, frontier models demonstrated performance variance spanning nearly two orders of magnitude across identical environments.\n\nComplete evaluations, research methodology and technical documentation are available at [olamlabs.ai/research/social-arena](https://olamlabs.ai/evaluations)\\\n\\\n**Launch**\\\n\\\nWe already passed a thousand matches a day in our early access, are nearing a billion tokens used per day, ranking on the OpenRouter leaderboards, and have thousands of users that are paying to play in the arena while we work with researchers on use cases they have for our work and environments. We’re hoping this is just a stepping stone and want to scale towards our roadmap and ambitions of what we believe multi-agent environments, when properly developed with the right stakeholders in the training/eval ecosystem, can achieve for AI progress.\\\n\\\n\\\n**What’s in it for you?**\n\n1. **Play.** It's free, and every match directly informs the evals. Bring friends, try to beat the models, it’s harder than you'd think. 🃏\n2. **Speak with researchers.** We love speaking with researchers interested in this space and it’s all that we think about; if that’s you please feel free to reach out to us at [founders@olamlabs.ai](mailto:founders@olamlabs.ai) or on \\\n \\\n 🫡\n\n---\n\n## About the team\n\n**Om Buddhdev (CEO)** — Previously a Product Engineer then promoted to Staff Engineer at AI Dungeon / Latitude working on AI infra, agents, and consumer platform work, left for Olam Labs. Before that he was among the best Valorant players in the world (top 25, 8x radiant), at 17 a large content creator on TikTok with 195k followers, and as a young teenager made Roblox games played millions of times, among other things in his past. More on Om at \n\n**Shreshth Sharma (CTO)** — Background in production data engineering: was a Data Scientist at RBC while doing a double degree at the University of Waterloo in CS and Math, dropped out of both for Olam Labs. He has a background in many data and statistics projects, alongside briefly being a top 50 Fifa player. More on Shreshth at \\\n\\\nOm and Shreshth have been friends since middle school, and Olam Labs was born out of their personal interest and desire to build out the entire stack for multi-agent environments to be the default way we evaluate models in the future.", "url": "https://www.ycombinator.com/launches/ST2-can-you-beat-frontier-llms-at-social-strategy-games-multi-agent-arena-by-olam-labs-evaluating-through-multi-agent-simulations", "votes": 9, "created_at": "2026-08-05T21:32:50.566Z" } ] }, { "id": 32131, "name": "Shielded", "slug": "shielded", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/eec883dcf84a87963c1d7874cef470c16c9dad19.png", "website": "https://www.shieldedglobal.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Shielded connects companies’ internal and external data to map market shifts, tariffs, and disruptions directly to suppliers, products, and margins. We turn those impacts into real-time decisions that grow and protect the business. ", "one_liner": "We help Fortune 500 companies hedge their supply chain risk", "team_size": 0, "industry": "B2B", "subindustry": "B2B -> Finance and Accounting", "launched_at": 1784606823, "tags": [ "Fintech", "Supply Chain", "Agriculture", "Energy", "Infrastructure" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Finance and Accounting" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/shielded", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/shielded.json", "year_founded": null, "city": "San Francisco", "country": "US", "linkedin_url": "https://linkedin.com/company/shieldedglobal", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/eec883dcf84a87963c1d7874cef470c16c9dad19.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [], "jobs": [], "news": [], "launches": [ { "id": 109233, "title": "Shielded: The company brain for supply chain risk", "tagline": "Shielded helps manufacturers understand and act on supply chain risk in real time.", "body": "Shielded helps manufacturers understand and act on supply chain risk in real time.\n\nAs the world has become more volatile, businesses need a tool to monitor events and rapidly take action. Founder James Simon saw this problem first hand in commodities trading, in fact his whole job was to take advantage of other firms reacting too slowly as the world changed. \n\nThat’s what inspired him to build Shielded, as an intelligence layer that integrates quickly, deeply understands your specific business, and turns volatility into clear, proactive decisions.\n\nShielded supports critical part and supplier management for leading defense companies, ingredient price audits for multinational CPG firms, and tariff impact analysis for global manufacturers.\n\nWith API, MCP, BOM, and spreadsheet integrations, every customer deployment to date has taken less than 72 hours.\n\nInstead of simply receiving an alert when something happens, Shielded gives procurement, finance, and supply chain teams company-specific guidance on the impacts of external events, market shifts, and regulation changes, and recommends the right operational or hedging responses.\n\nShielded is your supply chain’s command center that never sleeps.\n\n", "url": "https://www.ycombinator.com/launches/SPp-shielded-the-company-brain-for-supply-chain-risk", "votes": 5, "created_at": "2026-08-06T18:00:00.390Z" } ] }, { "id": 32133, "name": "Belvedir", "slug": "belvedir", "former_names": [ "Fractal", "Fractal Research", "Fractal Machine Research", "Belvedir", "Fractal Machine Research" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/7f68cfb5228c7e522deffed4be2be94da5eab6d3.png", "website": "https://belvedir.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "", "one_liner": "The autonomous AI model factory", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1784319754, "tags": [ "Machine Learning" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/belvedir", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/belvedir.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/belvedir/", "x_url": "https://x.com/belvedir_ai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/7f68cfb5228c7e522deffed4be2be94da5eab6d3.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 2576778, "full_name": "Zachary Yu", "title": "Founder/CEO", "founder_bio": "", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/zacharyyu/", "x_url": "https://x.com/thezacharyyu", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/cdfdf5e0aa8f760e0a5a920eaa5ad0f3b215ab37.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109999, "title": "Belvedir: Making private AI models cheap and easy so the world can have trillions of them", "tagline": "Make a custom AI model that’s 10x cheaper and 2x better with less than 5 mins of set up", "body": "Hey, I’m[ ](https://bookface.ycombinator.com/user/2576778)Zachary Yu, the founder of [Belvedir](https://belvedir.ai/). It’s nice to meet you!\n\n**TD;LR:** \n\nPrivate AI models are cheaper, better, faster, and protect your data. But right now they’re only used by big companies since they’re really difficult and expensive to make. Belvedir makes private AI models autonomously with less than 5 minutes of set up. By making it cheap and easy to make models, we hope people use Belvedir to create trillions of them.\n\nRight now, we’re live with 10 companies who are seeing up to 10x lower inference costs and 2x improved performance on benchmarks. \n\n \n\n**How it works:**\n\n1. Get an API key from [Belvedir ](https://belvedir.ai/)and install our SDK\n2. Belvedir turns your traces and logs into organized training sets and RL environments\n3. We autonomously train models from your data and optimize harnesses on top of these models\n4. We make and run benchmarks on these new models and harnesses\n5. We deploy these models and give you an API key to access them through \n6. We give you a router that gradually send more traffic to your new model as it out performs your old one\n7. As you use the new model we collect more traces and logs\n8. The loop repeats\n\nYour data is only used for you. Your models are only accessible by you. We use LoRA and can charge by the token by hosting multiple adapters on one cluster (so no worries if you have spikey traffic). \n\n**About Me:**\n\nI’m Zachary Yu, the founder of [Belvedir](https://belvedir.ai/). Before this, I founded [Traverse](https://www.traverse.so/): a YC startup that sells RL environments to frontier labs, studied CS at UWaterloo, and made RL environments at Mercor.\n\n**Long Term Ambitions:**\n\nWe want to make private AI the default for every use case. Starting with startups / SMBs and then consumers. \n\nWe are bullish on a world where there are trillions of private AI models. We want to build the ecosystem where these models are made, hosted, and interact with each other.\n\n**Asks:**\n\n1. If you have a high inference bill, want better performance from your agent, or want to protect your data from frontier labs, **we can help you out**.\n2. We are looking to hire engineers who have worked on RL / optimization, model inference, training infra, or just have that dog in them\n\nContact us at [founders@fractalresearch.ai](mailto:founders@fractalresearch.ai) ", "url": "https://www.ycombinator.com/launches/ScB-belvedir-making-private-ai-models-cheap-and-easy-so-the-world-can-have-trillions-of-them", "votes": 64, "created_at": "2026-08-11T22:00:00.240Z" } ] }, { "id": 32134, "name": "Parasma", "slug": "parasma", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/65a63b540239ee3a5538326d626bcceee19f52d9.png", "website": "https://parasma.com", "all_locations": "San Francisco, CA, USA", "long_description": "We train human brain cells for compute", "one_liner": "Replacing GPUs with human brain cells", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1782285260, "tags": [ "Machine Learning", "Neurotechnology", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/parasma", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/parasma.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/parasma", "x_url": "https://x.com/ParasmaAI", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/65a63b540239ee3a5538326d626bcceee19f52d9.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/adc99c68d73202b3a5001d44bbd5177d7a8335d0.png" ], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 3388159, "full_name": "Sean Cole", "title": "Founder/CEO", "founder_bio": "Founder at Parasma (YC S26)\n\nSkipped BS, MS in Artificial Intelligence w/ Distinction", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/sean-cole-8985a4207/", "x_url": "https://x.com/vytalow", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6f2d0bdfcf98a7f2843ea62be9a5893871caf196.jpg" } ], "jobs": [ { "id": 98189, "title": "Founding Scientist", "url": "https://www.ycombinator.com/companies/parasma/jobs/Vg7GVVl-founding-scientist", "location": "San Francisco, CA, US / Remote", "type": "Full-time", "role": "Science", "role_type": "Research", "salary_range": "$1M", "equity_range": "1.00% - 3.00%", "experience": "11+ years", "visa": "Will sponsor", "skills": [] } ], "news": [], "launches": [ { "id": 109992, "title": "Parasma: Training human brain cells for compute", "tagline": "In March, we trained them to play Doom; now we got them to do next-token prediction.", "body": "We train human brain cells for compute. In March, we trained them to play Doom; now we got them to do next-token prediction.\n\n![uploaded image](/media/?type=post&id=109992&key=user_uploads/3388159/cbefdb79-ad5b-4e03-ae35-d2a482a05351)\n\n![uploaded image](/media/?type=post&id=109992&key=user_uploads/3388159/ce53a90f-7180-4eab-b23e-d770a5e375a7)\n\n**The Problem**\n\nModern models require enormous amounts of energy and training data, while still struggling with continual learning and adapting efficiently to new tasks. Biological brains solve these problems very differently. They learn continuously, adapt from limited experience, and operate on roughly the power of a light bulb.\\\n\\\n**Solution**\n\nParasma is developing the algorithms and infrastructure needed to turn human brain cells into useful compute. The human brain is the most capable and energy-efficient learning system we know. Our goal is to reliably train biological substrates and use them for general computation, robotics, and eventually direct biological training. In March, we trained human brain cells to play Doom, which got hundreds of millions of views across social media. Since then, we’ve been working toward more general tasks. Most recently, we got the cells to perform token prediction: given a sequence of tokens converted to electrical stimuli, the neurons learn to predict what comes next with no GPU.\n\nLong-term, biological compute will offer orders-of-magnitude improvements in energy and sample efficiency, while enabling native continual learning.\n\n**Team**\n\nI previously worked on neuromorphic computing and reinforcement learning during my master’s.\n\nBefore that, I built a successful digital-item marketplace with > $1.7M of listed goods and played games professionally.\\\n\\\nCheck out our press release here:\\\n", "url": "https://www.ycombinator.com/launches/Sc4-parasma-training-human-brain-cells-for-compute", "votes": 10, "created_at": "2026-08-13T20:30:00.164Z" } ] }, { "id": 32135, "name": "Trope", "slug": "trope", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/1241956aa50c817e1c2feee9934ea253b0d12c38.png", "website": "https://trope.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Trope deploys custom AI agents into ERPs for mid-market companies. Every ERP is heavily customized, so useful agents need custom deployment too. Trope’s AI FDE learns each business and ERP, then builds and deploys agents, turning the last mile from services into software.", "one_liner": "AI FDE that deploys custom AI agents into ERPs", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1783447220, "tags": [ "SaaS", "B2B", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/trope", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/trope.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/tropeai", "x_url": "https://x.com/trope_ai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/1241956aa50c817e1c2feee9934ea253b0d12c38.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 2519723, "full_name": "Matthew Chow", "title": "Founder", "founder_bio": "Co-Founder at Trope. Safety critical firmware for the Tesla Cybertruck. Autonomous drone manufacturing at Zipline. Industrial software at Schneider Electric in Japan. UBC computer engineering grad.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/mdchow/", "x_url": "https://x.com/mattchowx", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/66b3684257cdd50e21184440e262dd2fe78b16e2.jpg" }, { "user_id": 1405446, "full_name": "Victor Vannara", "title": "Founder", "founder_bio": "Co-Founder at Trope. Built the first AI customer support agent at Amazon Payments. Learned to code at 13 by building Discord bots used by 400k users. UBC computer science grad.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/victor-vannara/", "x_url": "https://x.com/victorvannara", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d6b05e4463fd575d99cdde4b76ac50eb2f0fd14d.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 104941, "title": "Trope - Modernize ERP implementations with AI", "tagline": "The AI-native platform to deliver successful ERP implementations faster.", "body": "Hey everyone! We're [Matthew](https://www.linkedin.com/in/mdchow/) and [Victor](https://www.linkedin.com/in/victor-vannara/) - cofounders of [Trope](https://trope.ai/). We’re building the AI-native platform for ERP services.\n\n**TLDR**\n\n* After buying an ERP, companies spend over $50B a year on consulting services to get the software implemented\n* 70% of ERP implementations fail to meet budget, timelines, or targets\n* Trope deploys agents to help ERP implementation teams deliver projects on time and on budget from one AI-native platform\n\n**Launch Video:** \n\n\\\n**❌ Problem**\n\nCompanies buy an ERP expecting it to become the most powerful software in their stack. It runs finance, inventory, procurement, operations, reporting, and the workflows the business depends on.\n\nBuying an ERP is easy. Implementing it is hard and requires labour heavy consultant work where:\n\n* Requirements get missed\n* Workflows get configured differently than agreed\n* Data issues get discovered too late\n* Testing reveals the system does not match how the business actually works\n* End-user adoption suffers\n\nFor implementation teams, the challenge is managing hundreds of stakeholders across concurrent projects, with thousands of documents, decisions, and system dependencies to track.\n\nThat is why so many ERP projects miss their budget, timeline, and targets, and why companies spend billions on implementation services.\n\n**✅ Solution**\n\nTrope is the AI-native platform that deploys agents to deliver successful implementations faster.\n\nUsing Trope is like having a project manager, consultant, and developer working 24/7, so you can spend more time on the design and decisions.\n\nTrope connects to the systems where project work already lives. It compares what was promised with what is actually being built, then flags gaps across scope, configuration, data, testing, training, and documentation before they become expensive.\n\nTrope takes action by turning meetings into action items, scanning ERP configurations programmatically, and keeping documents, configurations, and stakeholders in sync.\n\n**The Team**\n\nWe grew up in Vancouver and met at the University of British Columbia. We built operations and AI software at Tesla, Amazon, Zipline, and Schneider Electric.\n\nAfter speaking with ERP veterans who had built and sold practices, the pattern was clear: everyone knew the traditional services model was going to change, but no one knew how.\n\nTrope is defining the AI-native model for ERP services.\n\n![uploaded image](/media/?type=post&id=104941&key=user_uploads/2519723/896d82bc-589c-4f7b-a0a7-59d2c9fc8d9f)\n\n**Our Ask**\n\nIf you know anyone who works on ERP implementations (ie, consulting, system integration, or professional services) we’d love an intro. Email me at [matthew@trope.ai](mailto:matthew@trope.ai).\n\nTake care,\n\nMatthew & Victor", "url": "https://www.ycombinator.com/launches/RIb-trope-modernize-erp-implementations-with-ai", "votes": 13, "created_at": "2026-07-08T17:03:27.360Z" } ] }, { "id": 32136, "name": "Palisade", "slug": "palisade", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/e033a9314af8c0fd0b02905c77d92b63b32820e2.png", "website": "https://palisade.run", "all_locations": "San Francisco, CA, USA", "long_description": "Palisade spins up a personal sales agent for every single visitor to your website. Marketplaces like trading card platforms, vintage stores, and short-term rental operators sell more when each customer works with a personal agent that remembers them.", "one_liner": "AI-native sales agents that run marketplaces", "team_size": 1, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1786633082, "tags": [ "Marketplace", "Sales", "E-commerce", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/palisade", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/palisade.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/palisader/", "x_url": "https://x.com/palisadelabs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/e033a9314af8c0fd0b02905c77d92b63b32820e2.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 1155886, "full_name": "Jonathan Salama", "title": "Founder/CEO", "founder_bio": "CEO & Founder of Palisade. Previously SWE @ Amazon Traffic & Marketing, Founding Eng @ Sage AI Labs, and co-creator of Lulus TV, a YouTube channel with 1M subscribers. CS @ Princeton.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jonathan-salama/", "x_url": "https://x.com/jonnysalama", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1ab0440756994e162a7f09a6aad6ada793b4683c.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110474, "title": "Palisade: Personalized sales agents for online marketplaces", "tagline": "We spin up a dedicated sales agent for every single visitor to your marketplace.", "body": "**Palisade builds agents to run your online marketplace. Our agents have driven over $2.7M in GMV since the start of the batch.**\n\n\n\n**The problem:** The best marketplaces in the real world run on salespeople. The real estate agent who takes you house to house and texts you the moment the right listing is about to hit the market. The Sephora beauty advisor who finds your exact shade. The jeweler who walks you through diamond cuts for an engagement ring. **Online marketplaces skipped that person entirely** and left the selling to search bars and filters.\n\n**What we do:** Palisade builds the missing **sales agent**. Every single visitor to your marketplace gets their own dedicated agent, and conversion goes up massively.\n\n**How it works:** Agents can take any action possible on your platform. They know every listing, every UI component, and every flow. They navigate for users with their own cursor, help them discover, add to cart, and complete the purchase or booking. They build custom pages and reshape the UI around how each person actually uses the platform, remembering each user and tailoring the experience to their behavior and preferences. Each dedicated agent can naturally text and email with their customer as well.\n\n**Is this another chatbot?** Each agent is proactive with much richer, customized components than an idle chatbot. It’s like giving every customer an omnipotent Claude embedded into the platform. Palisade agents get to know each customer personally and build relationships over time. They keep working after customers leave, remember them when they return, and reach out proactively over email and text with personal messages. And everything they learn flows back: agents aggregate their conversations into reports, insights, and recommendations for sellers, and offer to make concrete optimizations themselves, to the listings and even to the marketplace platform itself.\n\n**Integration**: Start by adding just one line of code to your existing platform, whether your marketplace sells yachts, vacation rentals, rare collectibles, or anything else.\n\n**Backstory:** After studying CS at Princeton, I spent years at Amazon building software for influencers to sell more products. I then became a founding engineer at Sage AI Labs, where I worked on some of the very first iMessage agents for consumer shopping. That work left me with a deep understanding of consumer habits and marketplace dynamics.\n\n**If you’re looking to transform your marketplace:** Reach out at [jonathan@runpalisade.com](mailto:jonathan@runpalisade.com)\n\n![uploaded image](/media/?type=post&id=110474&key=user_uploads/1155886/7772b584-f4ce-425c-93a3-eb36c96aaa46)\n\n2 days before the batch started. run palisade.", "url": "https://www.ycombinator.com/launches/Sjq-palisade-personalized-sales-agents-for-online-marketplaces", "votes": 84, "created_at": "2026-08-14T16:00:00.301Z" } ] }, { "id": 32137, "name": "Kara", "slug": "kara", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/1db7948d6861e68f3fa79c81509a04c4aea01614.png", "website": "https://karalabs.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Kara is building diamond semiconductors. \r\n\r\nWe begin with growing ultra-high purity diamond wafers for advanced hardware. Used in data centers, spacecrafts, RF systems, and quantum hardware, our diamond moves heat up to 5× better than copper, helping critical systems run cooler, faster, and more reliably. \r\n\r\nDiamond has spent centuries as a symbol. We are turning it into infrastructure.", "one_liner": "Making Diamond Semiconductors", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1785349021, "tags": [ "Hardware", "B2B", "Manufacturing", "Advanced Materials" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/kara", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/kara.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/kara-labs", "x_url": "https://x.com/KaraLabsAI", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/1db7948d6861e68f3fa79c81509a04c4aea01614.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3669602, "full_name": "Aarin Jhaveri", "title": "Founder/CEO", "founder_bio": "CEO @ Kara. 5th-generation diamond manufacturer. Sold millions worth of natural and lab-grown diamonds across the value chain. UC Berkeley Economics and Data Science. Industrializing diamond for the compute age.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aarin-jhaveri/", "x_url": "https://x.com/AarinJhaveri", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/bc3c1c5bffdd62cc5f29c634a07cf5bfca574fa3.jpg" }, { "user_id": 3669310, "full_name": "Anuveer Chadha", "title": "Founder", "founder_bio": "Founder @ Kara (S26) | Built diamond quantum sensing and computing systems @ LBNL with the Department of War | ML for chemical formulations @ Regeneron | Berkeley Chemical Engineering Dropout", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/anuveerchadha/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/11821dd27eb4615b23cd032eb75060954c633b4f.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109988, "title": "Kara: Making Diamond an Engineering Material", "tagline": "Kara grows and engineers diamond wafers for thermal management, semiconductor and quantum applications", "body": "**TL;DR:** We grow high-purity diamond wafers for scientific and industrial applications. Our diamond thermal wafers can be used in aircraft, spacecraft, RF electronics, and lasers to move heat dramatically better than conventional materials. We’re helping hardware run cooler, faster, and more reliably with diamond that is 5× more thermally conductive than copper.\n\nHere is a link to our launch video: \n\n\n\n**The Problem**\n\nThe world’s most valuable hardware systems are limited by how quickly they can remove heat.\n\nFor AI chips, that means expensive GPUs throttle and deliver less compute. In aircraft and satellites, it means carrying more cooling mass. In RF systems and lasers, it limits sustained output power.\n\nAdding more cooling costs space, and energy, and sometimes is simply not possible.\n\n**What We Built**\n\nDiamond is usually something you wear. We make it something engineers can build with.\n\nDiamond conducts heat up to **5× better than copper while weighing \\~60% less** for the same volume. It is electrically insulating and stable in extreme environments.\n\nToday we manufacture:\n\n* Diamond heat spreaders for AI and high-performance computing\n* Ultra-high-purity diamond wafers for thermal, photonic and quantum applications\n* Diamond-metal composites (copper and aluminium) for thermal management\n* Metallized and bondable diamond plates\n* Custom-engineered diamond components\n\nWe are now producing custom samples under pilot agreements and purchase orders for thermal, electronic, aerospace, and quantum applications.\n\nSilicon transformed computing when we learned to manufacture it with extreme purity and control. Kara is doing the same with diamond. Today, our ultra-high-purity diamond pulls heat away from silicon chips.\n\n**Tomorrow, Kara’s control over purity and defects unlocks diamond as the ultimate semiconductor.**\n\n**Why us**\n\n[@Aarin Jhaveri](https://bookface.ycombinator.com/user/3669602) is a fifth-generation diamantaire who has sold millions of dollars in natural and lab-grown diamonds and worked across the supply chain.\n\n[@Anuveer Chadha](https://bookface.ycombinator.com/user/3669310) is a chemical engineer who has worked on diamond quantum sensing and advanced-materials research at Lawrence Berkeley National Laboratory with the Department of Defense and Department of Energy.\n\nDiamond is our unfair advantage. We know how to grow, control, and engineer it far beyond jewelry.\n\n**Our ask**\n\n**Today, we're looking for introductions to engineers building hardware that is limited by heat.** \n\nSpecifically:\n\n* AI accelerator packaging teams\n* Advanced packaging / OSAT engineers\n* Aerospace, space, and defence electronics\n* Data-centre infrastructure companies\n* High-power RF and GaN teams\n\nIntroduce us to the engineer responsible for the hottest, hardest-to-cool hardware, especially where size, weight, or power makes cooling difficult. \n\nWe’ll build them a custom diamond sample to test.\n\nEmail us at [founders@karalabs.ai](mailto:founders@karalabs.ai) or book time here: \n\n**Diamond spent centuries as a symbol of “extraordinary”. We are making it something the world can build with.**", "url": "https://www.ycombinator.com/launches/Sc0-kara-making-diamond-an-engineering-material", "votes": 85, "created_at": "2026-08-11T20:00:00.166Z" } ] }, { "id": 32138, "name": "NebuSec", "slug": "nebusec", "former_names": [ "Nebula Security" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c60d9cd12e8c1b315e07ee45555c24b5f74d5654.png", "website": "https://nebusec.ai", "all_locations": "San Francisco, CA, USA; Remote", "long_description": "NebuSec is founded by the world’s #1 hacking team.\r\n\r\nWe were the first team to build an nginx RCE and the first to root Android 17. Our co-founders include members of the world’s No. 1 CTF team, DEF CON winners, Black Hat speakers, and PhD in cybersecurity. We have won $400K in bug bounties by exploiting the Linux kernel and Chrome browser.\r\n\r\n“Audited by NebuSec” is a mark of serious security credibility: a signal to customers, investors, and partners that your product has been reviewed by the world’s best hackers.", "one_liner": "Autonomous AI defense system for cyber attack", "team_size": 4, "industry": "B2B", "subindustry": "B2B -> Security", "launched_at": 1783007241, "tags": [ "B2B", "Security", "Cybersecurity", "Enterprise", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Security" ], "regions": [ "United States of America", "America / Canada", "Remote", "Fully Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/nebusec", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/nebusec.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/125323912", "x_url": "https://x.com/nebusecurity", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c60d9cd12e8c1b315e07ee45555c24b5f74d5654.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 1626290, "full_name": "Eten Zou", "title": "Founder/CEO", "founder_bio": "CEO & Founder @ Nebula Security (YC S26). PhD in Cybersecurity.\nFocus on automated vulnerability research.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/xiaochen-zou-phd-29540b168", "x_url": "https://x.com/etenal7", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/57bbc9606ef7588538536a132b2206290ce936f1.jpg" }, { "user_id": 3676420, "full_name": "Yuan Tan", "title": "Founder/CTO", "founder_bio": "Founder at Nebula Security (YC S26). Black Hat USA speaker. PhD dropout in System Security with experience in DARPA-funded fundamental research on program analysis.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/yuantan1/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/da759eb576807bd53522ad4e38bb45e06f4c0862.jpg" }, { "user_id": 3676409, "full_name": "Frank Wu", "title": "Founder/COO", "founder_bio": "Founder of Nebula Security (YC S26). Black Hat USA speaker and DEF CON finalist with r3kapig, the world's No. 1 hacking team in 2025. He led the DARPA AIxCC automated program repair team and shipped end-to-end RCE/LPE exploit chains on Chrome, Firefox, Linux, and Android. He pwned Google's kernelCTF multiple times with zero-days, and has published research at top academic venues such as USENIX Security and NDSS. Ph.D. dropout.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/yifan-wu-cs", "x_url": "https://x.com/FrankOverF1ow", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/54eba66a2435edef35054374d3e93dbb8d0af4ba.jpg" }, { "user_id": 3680004, "full_name": "Xiaochuan Yu", "title": "Founder/VP of Vulnerability Research", "founder_bio": "Founder at Nebula Security (YC S26). Hacked into iPhone, Chrome, Firefox, Safari, QEMU, and various things in your daily life. DEFCON finalist for 3 consecutive years. $200k+ Bug Bounty in Chrome. Core member of r3kapig, the #1 hacking team in the world in 2025, on CTFTime. M.S. from UC San Diego. Research published at USENIX Security, BlackHat USA. Ph.D. droped out. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/xiaochuan-yu-3103a5297", "x_url": "https://x.com/Nyaaaaa_ovo", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8c980e06818698012626c4e85e58e70351bf6003.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 106602, "title": "Nebula Security: Bringing beyond Mythos protection to your product", "tagline": "#enterprise #cybersecurity #security #pentest #b2b", "body": "**TL;DR:** AI has changed cybersecurity. Don't let the next HuggingFace incident happen to your product. Hire VEGA, your best security engineer — a cybersecurity agent built by the world's best hacking team. It brings the same capabilities used to uncover critical vulnerabilities in browsers, operating systems, and internet infrastructure to protect your product. [Book a demo](https://nebula-security.cal.com/etenz/30min?overlayCalendar=true) today or find us at \n\n**The Problem:** Cybersecurity has changed. When the attackers have AI, how do you prevent the next HuggingFace incident from happening to your product?\n\nWe are here at Nebula Security for one reason, we are building the **world’s best cybersecurity agent**, bringing beyond Mythos protection to your product. \n\n### AI makes you fast. We make you secure\n\nIntroducing **VEGA**, the best cybersecurity agent that is built by the world's best hackers. An agent that has discovered critical vulnerabilities in major infrastructure and earned over **$400K** in rewards from Google.\n\nWatch our [launch video](https://youtu.be/gOzg5M5DhXI)\n\nWith VEGA, you can:\n\n* Autonomously find bugs in your codebase\n* Monitor code changes 24/7\n* Analyze root causes and generate Proofs of Concept\n* Provide ready-to-use patches\n* Understand the overall code health\n\nYou focus on bringing your product to the world. We manage the security.\n\n# The world saw the proof\n\nWhile ~~other companies~~ create hypes, we show you the proof.\n\nVEGA discovered\n\n* [IonStack](https://x.com/nebusecurity/status/2069707520160227688?s=20): the world's first fullchain exploit to root Android 17.\n* [GhostLock](https://thehackernews.com/2026/07/15-year-old-ghostlock-flaw-enables-root.html): a 15-year-old Linux kernel vulnerability capable of compromising nearly all the Linux devices in the world.\n* [nginx-poolslip](https://x.com/nebusecurity/status/2057071579876753643?s=20) and nginx-quicburst: the world's first and second vulnerabilities capable of remotely control your Nginx servers\n* And 91 more CVEs from Linux, Windows, Chrome Browser, Firefox, Nginx, Wordpress, and the other software you use every day.\n\nWe found vulnerabilities in the world's most complex software. Let us protect your product too.\n\n# The epic hacker team\n\n* Eten (CEO): PhD in Cybersecurity; Pwn2Own winner; speaker at Linux Security Summit, Off-by-One, IEEE S&P, USENIX Security, and NDSS. Previously worked at Microsoft.\n* Yuan (CTO): Previous cybersecurity PhD; Black hat speaker; Bringing VEGA to secure Linux Kernel.\n* Xiaochuan (Research Lead): From the world's #1 hacking team; three-time DEFCON participant; Jailbreak iPhone; Black Hat speaker; Previously worked at Zellic.\n* Frank (Research Lead): From the world’s #1 hacking team; DEFCON participant; Black hat speaker. Built the world’s first Android 17 root exploit.\n\nWe have won over **$400,000** from Google by breaking into their safest system.\n\nLet us show you what we can find in your system. [Book a demo](https://nebula-security.cal.com/etenz/30min?overlayCalendar=true) today.\n\nFind us at | Email to [founders@nebusec.ai](mailto:founders@nebusec.ai)\n\n![uploaded image](/media/?type=post&id=106602&key=user_uploads/1626290/4c0c365e-7cc4-469f-9c1c-c04a82c4f1b7)\n\n", "url": "https://www.ycombinator.com/launches/RjO-nebula-security-bringing-beyond-mythos-protection-to-your-product", "votes": 12, "created_at": "2026-07-23T21:25:03.912Z" } ] }, { "id": 32154, "name": "Zomma", "slug": "zomma", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/019855257e473ab4938156b12e16e4a3214a2f51.png", "website": "https://zommalabs.com", "all_locations": "San Francisco, CA, USA", "long_description": "Zomma is building the AI workforce for financial services with computer-use agents that onboard like humans and scale like compute. Its agents handle KYC reviews, transaction alerts, and disputes on the same screens human analysts use, no integration.", "one_liner": "The AI BPO for financial servies, run by computer-use agents", "team_size": 5, "industry": "B2B", "subindustry": "B2B", "launched_at": 1783572526, "tags": [ "Robotic Process Automation", "Finance", "B2B", "Workflow Automation", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/zomma", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/zomma.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/zomma", "x_url": "https://x.com/ZommaLabs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/019855257e473ab4938156b12e16e4a3214a2f51.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/b2b79e62b805cf93e198d9b3d4b6342664e73d7c.png" ], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3128623, "full_name": "Jihyun Kim", "title": "Founder/CEO", "founder_bio": "Founder at Zomma (YC S26). Launched $1B Hedge Fund as the Founding Team and Head of Product. Scaled Asia's largest wealth tech (now $10B+ AUM). Second time founder. Ex-Goldman Sachs. Cornell and NYU Abu Dhabi.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/kimjihyun/", "x_url": "https://x.com/jihprobs", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/67617ac3a534465e42aa00e5e395beaed235f970.jpg" }, { "user_id": 3597712, "full_name": "Varun Arumugam", "title": "Founder", "founder_bio": "Founder at Zomma (YC S26). Previously Founding ML Engineer at Bronco AI building agents for chip design verification. SWE at Apple on verification infrastructure. MS and BS in Electrical and Computer Engineering from UT Austin.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/varunarumugam/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/526b6e3a7761d998cfb1049aab57c89fa1f3cd11.jpg" }, { "user_id": 3627790, "full_name": "Cashel Fitzgerald", "title": "Founder/CTO", "founder_bio": "Co-founder at Zomma. Master's in Computer Science from Cornell. Built SOTA aerial object detection systems at Modern Intelligence. Previously engineered backend systems at 8am supporting $300M+ annual revenue and $20B+ yearly transactions. Trains VLMs for computer use.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/cashelfitzgerald", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/69c3f4c9b45abfcd2073a8a53036f40eba0a1ce1.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110619, "title": "Zomma: AI operations employees for finance back offices", "tagline": "Computer-use agents that work cross-border transaction alert investigations end to end: collect evidence, produce an assessment, draft an RFI.", "body": "We're building Zomma: AI operations employees for finance back offices, starting with cross-border transaction alert investigations.\n\n**TL;DR:** Zomma agents work directly inside the software your ops team already uses. No APIs, no integrations. For every cross-border alert, the agent gathers the evidence, writes the draft assessment, and drafts the RFI, end-to-end. Running live compliance-ops pilots with a US RIA and a $10B AUM wealth tech. 92% on OSWorld.\n\n\n\n## **The problem**\n\nCross-border payments trip more monitoring rules than anything else in the queue: sanctions name hits on foreign counterparties, high-risk corridors, first-time beneficiaries. Most alerts are false positives (McKinsey: over 90% in transaction monitoring), but every one still needs to be investigated. 20 to 60 minutes per case, across five or six systems, several with no API at all. And the evidence that clears them sits in foreign company registries and local-language adverse media. \n\nSo review capacity scales the only way it can: hiring. When hiring lags volume, alerts pile into backlogs that take months to clear, and backlogs are what regulators fine: $3.8B in AML penalties last year alone.\n\n## **What Zomma does**\n\nZomma agents work the queue on the same screens your analysts use: case manager, screening tool, admin panel, inbox, web search. Onboarding looks like hiring: an SOP and a login.\n\nOn a transaction monitoring alert, the agent\n\n* pulls the KYC profile and payment history\n* verifies the invoice and researches the counterparty, including registries you can only search by hand\n* writes a case-specific assessment with every claim cited to the source screen\n* drafts the RFI when documents are missing\n\nEvery step is logged and replayable: timestamped screenshots of every page checked, numbered citations, explicit no-match statements, a screen-by-screen trail an examiner can follow.\n\n## **Why start with cross-border alerts**\n\nIf an employee can do it on a screen, Zomma can learn it: the rest of the compliance queue, payment disputes, client onboarding, trade ops. Cross-border alerts come first because they have the exact shape our agents are best at: high-volume, parallelizable work that needs real reasoning but follows a defined workflow to a documented disposition. They're also the hardest cases to automate any other way: the evidence lives in registries and portals that will never have APIs. And agents scale with the queue: a volume spike means provisioning two more that afternoon; a BPO batch takes a month to train.\n\nComputer-use agents only recently got accurate enough to trust with a live queue, which is why we're building this now. Our state-of-the-art computer-use harness with our procedural and workflow memory system scores 92% on OSWorld.\n\n## **Team**\n\n* **Jihyun (CEO):** founding team at a Blackstone-backed hedge fund, scaled Asia's largest wealth tech, Goldman engineer.\n* **Cash (CTO):** backend systems behind $20B+ in yearly transactions, trains VLMs for computer use.\n* **Varun (CPO):** high-performance infra at Apple, design-verification agents at Bronco AI (YC S23).\n\nSend us \\~100 closed cross-border alerts and we'll show you what the agent would have done with them, or book a call: zommalabs.com/call.", "url": "https://www.ycombinator.com/launches/SmB-zomma-ai-operations-employees-for-finance-back-offices", "votes": 76, "created_at": "2026-08-18T16:00:00.082Z" } ] }, { "id": 32155, "name": "Rex", "slug": "rex-inc", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/b14c16514ac7654483b96e54e1bf49a45714e8f6.png", "website": "https://www.rex.inc", "all_locations": "San Francisco, CA, USA", "long_description": "Rex is an AI-native service for order to cash. Rex replaces manual operations with AI agents that continuously manage invoices, customers and exceptions.\r\n\r\nOrder to cash is a $60B outsourced function because the work is too messy for traditional automation. Every customer has different context, exceptions do not follow clean rules, and the next step is often buried in emails, portals, handoffs, or someone’s memory. Rex agents run continuously in the background and resolve work to SLA standards so cash keeps moving. We're live with pre-IPO tech companies, managing >$500m in receivables from F100 companies, cutting manual follow-up cycles from weeks to hours.\r\n\r\nWe founded Rex after spending four years building Sequence, an a16z-backed finance ops platform, from zero to Series A — where we saw firsthand that at-scale, O2C still runs on manual labour.\r\n\r\nRex is the #1 winner of the Vercel AI Accelerator out of 2,500 teams, selected by Guillermo Rauch and leaders from OpenAI, Anthropic, Browserbase and more.", "one_liner": "AI-native BPO for enterprise order-to-cash", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Finance and Accounting", "launched_at": 1779903831, "tags": [ "B2B", "Enterprise", "Enterprise Software", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Finance and Accounting" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/rex-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/rex-inc.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/rex-ai-inc/", "x_url": "https://x.com/rexdotinc", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/b14c16514ac7654483b96e54e1bf49a45714e8f6.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/c5e473e7845a96325de3a6590b9157c80ee928a7.png", "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/242d907a5de4ed9944a88c07e410d6a12acb5cb2.png" ], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 466573, "full_name": "Merlin Kafka", "title": "Founder/CEO", "founder_bio": "Merlin is the Founder and CEO of Rex, the AI partner for enterprise order-to-cash.\n\nHe spent 4 years as CPO at Sequence (a16z-backed), an order-to-cash platform where he grew the company to 100+ customers. He has deep domain expertise in order to cash and accounts receivable workflows -- the exact space Rex is transforming with AI.\n\nPreviously a PM at Tessian (Sequoia-backed, acquired in 2023). Studied Industrial Engineering in Cambridge and Product Design in Glasgow.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/merlinkafka/", "x_url": "http://x.com/merlinkafka", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/47ef8bf61952f3ecae050d622d963369e205daa4.jpg" }, { "user_id": 173279, "full_name": "Lewis Blackwood", "title": "Founder/CTO", "founder_bio": "Founder and CTO at Rex, building the AI partner for enterprise order to cash.\n\nPreviously:\n- Founding engineer at Sequence (backed by a16z)\n- Cofounded Personably (backed by Global Founders Capital)\n- Early team at GoCardless (YC S11)\n- Economics at Cambridge", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/lewisblackwood", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2619bd1bf4df714f4e821b2c1ba6a7c28caafede.jpg" } ], "jobs": [ { "id": 113085, "title": "Founding AI Engineer (Enterprise Finance Operations)", "url": "https://www.ycombinator.com/companies/rex-inc/jobs/z3eroAv-founding-ai-engineer-enterprise-finance-operations", "location": "London, England, GB", "type": "Full-time", "role": "Engineering", "role_type": "AI engineer", "salary_range": "$150K - $250K", "equity_range": "0.10% - 1.00%", "experience": "1+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 113092, "title": "Finance Operator in Residence (Order-to-cash)", "url": "https://www.ycombinator.com/companies/rex-inc/jobs/87h0X9A-finance-operator-in-residence-order-to-cash", "location": "London, England, GB", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "$60K - $120K", "equity_range": "0.10% - 0.50%", "experience": "1+ years", "visa": "US citizenship/visa not required", "skills": [] } ], "news": [ { "title": "Rex in Forbes | What Y Combinator's Latest Batch Reveals About The Future", "url": "https://www.forbes.com/sites/dariashunina/2026/06/04/what-y-combinators-latest-batch-reveals-about-the-future/", "date": "Sep 03, 2026" }, { "title": "How Rex automates order-to-cash for finance teams", "url": "https://www.braintrust.dev/customers/rex", "date": "Aug 26, 2026" }, { "title": "How Rex went from zero to enterprise ready in weeks with WorkOS", "url": "https://workos.com/blog/how-rex-went-from-zero-to-enterprise-ready-in-weeks", "date": "Jul 01, 2026" }, { "title": "Rex wins Vercel's AI Accelerator out of 2500 teams", "url": "https://vercel.com/blog/2026-vercel-ai-accelerator-recap", "date": "Apr 28, 2026" } ], "launches": [ { "id": 107722, "title": "Rex: Turn enterprise order-to-cash into an autonomous operation", "tagline": "AI Operations Partner for Enterprise Order to Cash", "body": "Rex is an AI service that runs order-to-cash and accounts receivable operations for enterprises. Founders [Merlin Kafka](https://www.linkedin.com/in/merlinkafka/) and [Lewis Blackwood](https://www.linkedin.com/in/lewisblackwood) spent 4 years at an a16z-backed startup helping finance teams at startups like Cognition, Legora and Bridge automate revenue collection. The problem is 100x bigger in the enterprise, so they built Rex: agents that do the work end-to-end. Live in production with Synthesia.\n\n\n\n**Revenue ≠ cash**\n\nEnterprises invoice customers, then spend months chasing the money through email chains, PDFs and ancient web portals. A $1M invoice sits unpaid because one PO number didn't match.\n\nMultiply that across thousands of customers and you get 100+ person teams, often offshore, doing this by hand: collections, invoice submission, dispute resolution, deduction management, cash application. Order-to-cash is the most outsourced process in finance because the work needs judgment. Reading a customer's email and deciding whether their dispute is valid. Working out which of 14 open invoices a lump-sum payment covers. Knowing that this account always pays late but always pays, and that one needs escalating now.\n\nAR automation software has existed for 20 years and the teams never got smaller, because the tools organise the work while people still do it. Businesses invoice each other tens of trillions of dollars a year; the labour and outsourcing spend on collecting it is over $100B. Until now the only way to scale was headcount or a BPO contract.\n\n**What Rex does**\n\nRex connects to the ERP (NetSuite, SAP, Oracle), CRM (Salesforce), billing system and shared AR inbox, and learns how the team runs the process today. That knowledge becomes editable agent operating procedures: how to chase a Fortune 500 buyer vs a startup, which deductions get investigated vs written off, when a dispute goes to a human.\n\nThen agents run those procedures across thousands of accounts in parallel:\n\n* Collections and payment follow-ups, including the long tail no team gets to\n* Invoice submission and status tracking inside procurement portals (SAP Ariba, Coupa, Tungsten and hundreds of others)\n* Dispute and deduction investigation\n* Remittance processing and cash application\n* Responding to customer finance emails\n* Updating ERP, CRM and billing records\n* Routing exceptions to the team for approval\n\nPortals deserve a special mention: an increasing share of enterprise invoices must be submitted and tracked inside the buyer's procurement system. It's the work AR teams hate most and the work software vendors ignore. Rex treats it as core.\n\n**Is Rex a managed service or software?**\n\nEnterprises can partner with Rex as a managed service provider or let their team operate Rex. AR software (dunning tools, collections worklists, AR dashboards) organises work for a team to do. Rex agents do the work, and the finance team approves exceptions. With Rex, the team stops drowning in routine chasing and handles the judgment calls. Customers compare Rex to their next AR hire, their offshore team, or their BPO contract.\n\n**Can finance teams trust AI agents with collections?**\n\nAutonomy is earned, not switched on. Each procedure starts supervised: Rex proposes, the team approves. Once a procedure hits its accuracy bar for the account, it graduates and runs autonomously. Safe, high-volume work goes first; judgment-heavy edges stay gated as long as the team wants. Every action is logged with its full reasoning, so finance audits the system instead of babysitting it.\n\n**Traction**\n\n[Rex is live in production with enterprises like Synthesia](https://www.rex.inc/customers/synthesia), running collections, portal submissions, inbox triage and cash application. Cases that used to take weeks of back and forth resolve in minutes.\n\n**Get in touch**\n\nIf invoice volume is outgrowing the finance team, or order-to-cash runs on offshore headcount or a BPO contract, get in touch: [rex.inc](http://rex.inc)\n\n![uploaded image](/media/?type=post&id=107722&key=user_uploads/466573/80aa1a0e-383d-43f5-93b4-4a97d086d306)\n\n", "url": "https://www.ycombinator.com/launches/S1S-rex-turn-enterprise-order-to-cash-into-an-autonomous-operation", "votes": 11, "created_at": "2026-07-27T18:08:15.052Z" } ] }, { "id": 32156, "name": "Enact", "slug": "enact", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/6f79f9bbc22f20d6e3465ca66081c4a646834064.png", "website": "https://enact.company", "all_locations": "San Francisco, CA, USA", "long_description": "Enact creates the physical environments to benchmark and hill climb robotics models on high-value tasks. We find where robots fail, and fix them with our in-house training loop.", "one_liner": "Real-world RL environments for robotics models", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1786311090, "tags": [ "Reinforcement Learning", "Robotics", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/enact", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/enact.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/enact-company/?viewAsMember=true", "x_url": "https://x.com/enact_labs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/6f79f9bbc22f20d6e3465ca66081c4a646834064.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 2406199, "full_name": "James Stevens", "title": "Founder/CEO", "founder_bio": "Founder & CEO @ Enact", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/james-stevens-440362200", "x_url": "https://x.com/jamesw_stevens", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c605b46207ec2b29443db68751da93ca45c507c0.jpg" }, { "user_id": 3594727, "full_name": "Govind Chada", "title": "Founder/CTO", "founder_bio": "Founder at Enact. Previously: Robotics research in Chelsea Finn’s IRIS Lab at Stanford AI Lab, Meta Reality Labs.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/govind-chada", "x_url": "https://x.com/govindchada", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/bffdb6ffe1344f591c02771b2dceccd816fe305e.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109880, "title": "Enact: post-training infrastructure for robotics models", "tagline": "Enact generates the data and evaluations needed to make robotics models reliable in the real world.", "body": "**TL;DR: Enact helps teams deploy robots that work in the real world. We roll out policies, generate failure / recovery data, and run in-house evaluations so robots can complete real tasks in production.** \n\n[**https://www.youtube.com/watch?v=dam0tQBIezY**](https://www.youtube.com/watch?v=dam0tQBIezY)\n\n**The Problem**\n\nRobotics models can look flawless in choreographed demos and still fail when an object slips, the scene shifts, or the robot makes an early mistake. Those failures push the policy into states its training data never taught it to recover from. **Robotics models must train on edge cases to be reliable in the real world.**\n\n**What We Built**\n\nEnact turns a policy’s failures into targeted recovery datasets for teams training robotics models. We use the data to fine-tune models and run real-world evaluations, verifying our data improves reliability.\n\n**We’re already serving our first customers: finding where their robots fail and collecting the targeted recovery data needed to make them robust.**\n\n* Roll out an existing model on a physical robot.\n* Identify recurring failure states.\n* Recreate those states and demonstrate how to recover.\n* Retrain with the new trajectories added in.\n* Repeat\n\n**Targeted recovery data pushes success rates above 99%.**\n\nOn our controlled packing task using pi0.5 on the YAM 6DoF arms:\n\n* 200 ordinary expert demonstrations: 90/100 successful physical rollouts\n* 250 ordinary expert demonstrations (more of the same data): 89/100\n* 200 expert demonstrations plus 50 targeted recovery demonstrations: 99/100\n\n![uploaded image](/media/?type=post&id=109880&key=user_uploads/2406199/e1b83c19-6e1d-4dda-adcf-258398bc1676)\n\n**We’re continuing to longer-horizon tasks such as industrial kitting, lab automation, and electronics recycling**\n\nLonger-horizon tasks need substantially more base and recovery data, but the loop is the same: find failures, target missing behaviors, retrain, repeat.\n\n**Who We Are**\n\nWe’re James Stevens and Govind Chada, and we met at Stanford. Govind researched real-world robot adaptation in Chelsea Finn’s IRIS Lab and co-developed recovery methods published at top robotics conferences. Together, we built Enact’s in-house data and evaluation operation and are already serving robotics teams. We started Enact to make robotics models work reliably outside the lab.\n\n**Our Ask**\n\nTraining a robot foundation model or deploying a policy that fails on a real task? Tell us the model, task, and where it breaks, and we’ll make it work.\n\nEmail: [founders@enact.company](mailto:founders@enact.company)", "url": "https://www.ycombinator.com/launches/SaG-enact-post-training-infrastructure-for-robotics-models", "votes": 16, "created_at": "2026-08-11T14:30:00.204Z" } ] }, { "id": 32157, "name": "Maingen", "slug": "maingen", "former_names": [ "Maingen AI", "Maingen", "Maingen AI" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/2eb59fa1eb6f59fafb938607c641276931a20aef.png", "website": "https://www.maingen.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Maingen builds simulations for industrial operations ($5 trillion slice of the US economy) so frontier labs can train models that actually run power plants, factories, and datacenters.", "one_liner": "Simulating Agents Running Industrial Companies", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Energy", "launched_at": 1784688239, "tags": [ "B2B", "Energy", "Big Data", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Energy" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/maingen", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/maingen.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/maingen", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/2eb59fa1eb6f59fafb938607c641276931a20aef.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 1925382, "full_name": "Phillip Yan", "title": "Founder/CEO", "founder_bio": "CEO at Maingen (S26). Prev. 0x, power trading, Scale AI, Coinbase (YC S12). \nShipped exchange infra powering $150M+ TPV", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/pyan/", "x_url": "https://x.com/PhillipYan", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ecf7bdd1531c6e539582b01d63661a59e9e880c5.jpg" }, { "user_id": 2362549, "full_name": "David Yang", "title": "Founder/CTO", "founder_bio": "CTO of Maingen (S26). Prev. Idler (S25), Scale AI, Reticular (F24), and Amazon.\nCited by Anthropic for ICLR2025 mech-interp paper.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/david-yang-ai/", "x_url": "https://x.com/djyang1", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4a0c350ff34675efa324bcf4a043c0999f005df3.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109472, "title": "Maingen: Simulations of real industrial operations, so agents can run the physical world", "tagline": "Long horizon RL environments for the $5 trillion industrial economy", "body": "Hello everyone!\n\nPhillip and David here, founders of Maingen (YC S26).\n\n**TLDR: AI is moving to the industrial world. Maingen builds RL environments for industrials (a $5 trillion slice of the US economy) so frontier labs can train models that actually run factories.**\n\n**Problem**:\n\nWe believe the next capabilities frontier is the physical world: operating factories, power plants, and industrial infrastructure. But that data doesn't exist on the internet, but rather as tacit knowledge inside enterprises, on operations desks, and in CMMS systems and vendor portals.\n\n**What we do:**\n\n\n\nWe embed with operators in these industries and translate their decision making process into RL environments labs can train.\n\n**Our first public benchmark evaluates whether agents can run a solar portfolio [solarbench.maingen.ai](http://solarbench.maingen.ai).** We’re continuing in other areas like factories, and we’re already working with frontier labs.\n\n**Who we are**\n\nThe founders Phillip Yan and David Yang have previously built RL environments, worked in the energy industry, shipped software powering 150 million+ TPV, and have been cited by Anthropic.\n\n**Get Involved**\n\nMaingen is looking for researchers and industrial operators to help shape its next environments. We’d love to hear from you if you:\n\n1. Work on research on long-horizon **agents, RL environments, or physical-world capabilities**.\n2. Run operations at an industrial site or can introduce the team to someone who does\n\nReach out at [founders@maingen.ai](mailto:founders@maingen.ai)! We're always down for a coffee!\n\nThanks,\n\nPhillip and David", "url": "https://www.ycombinator.com/launches/STg-maingen-simulations-of-real-industrial-operations-so-agents-can-run-the-physical-world", "votes": 5, "created_at": "2026-08-07T16:00:00.073Z" } ] }, { "id": 32160, "name": "MusicArena", "slug": "musicarena", "former_names": [ "Origami Research", "Catapult" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/97ef82ddda796191842e9ee0fb976e3ccf939dca.png", "website": "https://musicarena.io/", "all_locations": "San Francisco, CA, USA", "long_description": "https://musicarena.io/", "one_liner": "The crowdsourced benchmark for AI-generated music and music production", "team_size": -1, "industry": "B2B", "subindustry": "B2B", "launched_at": 1790372687, "tags": [ "Music", "Crowdsourcing", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/musicarena", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/musicarena.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/trycatapult", "x_url": "https://x.com/PlayCatapult", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/97ef82ddda796191842e9ee0fb976e3ccf939dca.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3041908, "full_name": "Kevin Yao", "title": "Founder/CEO", "founder_bio": "Co-Founder at Catapult (YC S26). UCLA dropout. Previously ML and robotics research at Berkeley & UCLA. CoRL 2026, AIME, USNCO National Qualifier, Science Olympiad medalist.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/kevinyao05/", "x_url": "https://x.com/truekyt", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6b8f9e448bffb8c0d943b9076a1772de16fd6644.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 117104, "title": "🎵 MusicArena - the arena for AI music-making 🎵", "tagline": " The crowdsourced benchmark for AI systems that make music", "body": "**Paucis verbis:**\n\nThe idea is simple: give two systems the same music task, listen to what they make, and vote for the better result.\n\nWe’re starting with AI agents producing music inside a DAW.\n\nGive two agents the same task, watch them work, listen to the finished tracks, and vote for the winner.\n\n\n\n**Why start with LLMs?**\n\nMost existing AI music benchmarks focus on models that generate audio end-to-end.\n\nWe think there’s another category that’s going to become increasingly important: general-purpose models that can hear music, reason about it, and operate the same creative tools humans use.\n\nFrontier models are already becoming much better at computer use, and we expect audio to increasingly become a native modality as well.\n\nToday, these agents are largely operating through tool calls/computer use rather than actually listening to their own output. But as audio becomes a native modality for frontier models, we expect agents to increasingly be able to hear the music they’re making and iterate on it directly.\n\nMusic makes for an especially interesting test because there often isn’t an objective answer, so human preference matters.\n\n**WE NEED YOU!**\n\nPull up to our website and try it out yourself! [musicarena.io](http://musicarena.io)\n\nWant your model on the leaderboard? Forward us an email at [founders@musicarena.io](mailto:founders@musicarena.io) 🦁", "url": "https://www.ycombinator.com/launches/USm-musicarena-the-arena-for-ai-music-making", "votes": 5, "created_at": "2026-09-28T14:55:16.503Z" } ] }, { "id": 32161, "name": "Bernard", "slug": "bernard", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/86e6151a4772f82f25d8581efbda66f538ce4bed.png", "website": "https://bernardhq.com/", "all_locations": "New York City, NY, USA", "long_description": "Bernard builds AI employees for home and commercial appliance servicers. They work across phone, text, email, warranty feeds, and existing systems to manage customer communication, scheduling, and dispatch, carrying each job through diagnosis, parts preparation, technician handoff, and follow-up. At the center is Pre-ID, which combines the exact appliance and its symptoms with repair history, similar completed visits, manufacturer guidance, and parts data to produce a diagnosis, ranked parts, and follow-up questions. Technicians arrive prepared with the right parts, helping servicers complete more repairs on the first visit.\r\n\r\nWe’re building the world’s most efficient appliance servicers, creating happier customers and a more connected appliance ecosystem.\r\n\r\nLearn more at https://bernardhq.com/", "one_liner": "AI employees that run home appliance repair companies", "team_size": 9, "industry": "B2B", "subindustry": "B2B -> Operations", "launched_at": 1786664902, "tags": [ "Home Services", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Operations" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/bernard", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/bernard.json", "year_founded": 2026, "city": "New York City", "country": "US", "linkedin_url": "https://www.linkedin.com/company/bernard-labs", "x_url": "https://x.com/BernardLabsHQ", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/86e6151a4772f82f25d8581efbda66f538ce4bed.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3076324, "full_name": "Dan Katzman", "title": "Founder/CEO", "founder_bio": "Dan is the co-founder and CEO of Bernard, where he focuses on building AI employees for the world’s most efficient appliance repair servicers. Before Bernard, he led strategy and operations at Nava Benefits and began his career at City Light Capital and Golub Capital. A native New Yorker, Dan graduated from the University of Southern California.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dankatzman/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/7e1bec6c246ca26e0bc9233f97856222281f0144.jpg" }, { "user_id": 3189022, "full_name": "David Goodfellow", "title": "Founder/CTO", "founder_bio": "Co-Founder at Bernard (YC S26). Previously Engineering Manager at Radar Labs, where he led the fraud and geo-compliance products. Started career at Disney+. USC B.S. CS/BA + M.S. Data Informatics.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dgoodfellow/", "x_url": "https://x.com/davothedev", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9243f8fd38407d2c55286323424aeb424c28d236.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111388, "title": "Bernard: AI front and back office for appliance repair companies", "tagline": "Run every repair from first call to first-time fix.", "body": "**Bernard is building the AI front and back office for home and commercial appliance servicers.**\n\nThey manage each repair from the first customer call through scheduling, dispatch, diagnosis, parts preparation, technician handoff, and follow-up. Bernard works across phone, text, email, warranty feeds, and existing field-service systems, carrying the job context across every surface.\n\nWe are already live in the field with some of the largest authorized service companies in the country.\n\nOur goal is simple: make the first visit the last one.\\\n\\\n\n\n### The first truck roll should not be a diagnostic visit\n\nAppliance repair is an information and inventory problem. A technician cannot simply carry washer parts. The exact appliance, model, revision, symptoms, and failure mode determine which parts might be required, and many nearly identical parts are not interchangeable.\n\nWhen that information does not come together before the appointment, the first visit becomes a diagnostic visit. The technician identifies the failure, orders the part, and returns days later to complete the repair. Customers wait longer, technicians complete fewer jobs, and servicers lose capacity and margin.\n\n### Pre-ID prepares the repair before the truck rolls\n\nWhen a customer reports that their washer will not drain, Bernard does more than book an appointment.\n\nBernard identifies the exact appliance, asks follow-up questions specific to the symptoms, and collects missing evidence from the customer. Pre-ID then combines that information with repair history, similar completed visits, manufacturer guidance, parts data, and the servicer’s own operating knowledge.\n\nBefore the visit, Bernard gives the technician a diagnosis, ranked parts, the most useful next questions, and the evidence supporting each recommendation. The office can prepare the parts, confirm they will arrive on time, and hand the technician a complete picture of the job.\n\nEvery interaction and completed repair feeds back into Bernard, turning the servicer’s daily work into shared knowledge.\n\n### Why start with the servicer?\n\nEvery appliance company makes a promise at the point of sale. The servicer is the person who ultimately has to keep that promise inside someone’s home or business.\n\nThat visit affects the customer, the retailer, the manufacturer, the warranty company, and the parts distributor. Yet the servicer coordinating it is often working across disconnected phone systems, inboxes, scheduling tools, parts catalogs, and knowledge held by a few experienced employees.\n\nWe believe that building the world’s most efficient appliance servicers will create happier customers and a more connected appliance ecosystem.\n\nBernard starts with the servicer because that is where the industry meets the physical world.\n\n### About us\n\n[Dan Katzman](https://www.linkedin.com/in/dankatzman/) previously helped build Nava Benefits' strategy and operations team, working across the product and operational infrastructure required to deliver a complicated, service-heavy experience reliably, while also leading partnerships. Before Nava, he invested at City Light Capital and Golub Capital.\n\n[David Goodfellow](https://www.linkedin.com/in/dgoodfellow/) grew from product engineer to engineering manager at Radar Labs, where he led fraud and geo-compliance products and helped replace the incumbent provider at DraftKings and Bet365. Before that, he worked on RPA and cognitive automation at Disney, before enterprise automation was cool.\n\n### Talk to us\n\nIf you run an appliance service company or lead service operations at an appliance manufacturer, retailer, warranty company, or service network, reach out to us at [founders@bernardhq.com](mailto:founders@bernardhq.com)\n\n![uploaded image](/media/?type=post&id=111388&key=user_uploads/3076324/2cf911ba-3755-4320-a8b0-e809cf265e3c)\n\n", "url": "https://www.ycombinator.com/launches/Sya-bernard-ai-front-and-back-office-for-appliance-repair-companies", "votes": 5, "created_at": "2026-08-20T20:30:00.201Z" } ] }, { "id": 32162, "name": "Studio", "slug": "studio", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/be382958698206c30f9e2a1996b85d6729c8ed0b.png", "website": "https://trystudio.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Studio is a simulation lab studying how humans and societies respond to new ideas, and is building the first software models that can accurately predict that response. Instead of creating digital twins of survey respondents, Studio builds live market models of interconnected audience segments from real, traceable consumer-behavior data. Our simulations match live KPI outcomes like revenue and purchase behavior with over 97% accuracy for consumer-facing enterprises, higher than any synthetic panel, simulation lab, or traditional market research.", "one_liner": "Accurately simulate market response at scale", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Analytics", "launched_at": 1787114400, "tags": [ "B2B", "Analytics", "Market Research", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Analytics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/studio", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/studio.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/trystudioai/", "x_url": "https://x.com/trystudioai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/be382958698206c30f9e2a1996b85d6729c8ed0b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 3118770, "full_name": "Nikita Mullangi", "title": "Founder/CEO", "founder_bio": "Founder at Studio. Researched human behavior & decision making @ MIT. Ex-UC Berkeley CS.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/nikita-mullangi/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/21d8a2a9ca356c4067f8c184faaebcaf75a61f29.jpg" }, { "user_id": 3141318, "full_name": "Cameron Malloy", "title": "Founder/CTO", "founder_bio": "Founder at Studio (YC S26). Built AI agents at GI Partners ($49B AUM PE firm). Created agent-based financial contagion models at Nelumbium Capital. Ex UC Berkeley M.E.T. (EECS + Business Admin)", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/cammalloy/", "x_url": "https://x.com/CameronMalloy_", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/de0981b075b9ead121a23ebff63588cc572203a9.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111412, "title": "Studio: The simulation engine for market behavior", "tagline": "The first accurate simulation models able to predict society's response to an idea.", "body": "**TL;DR:** Studio is a simulation lab studying how humans and societies interact with new ideas, and building the first software models that can accurately predict that response. Instead of creating digital twins of survey respondents, Studio builds live market models of interconnected consumer segments from real behavior data. Studio's predictions match live KPI outcomes like revenue and sentiment with over 95% accuracy.\n\n\n\n## The problem\n\nA huge number of consequential decisions are bets on human behavior.\n\nA company launches a product or changes a price. A brand runs a campaign. A public institution launches a new initiative or communications strategy. In each case, the outcome depends on how a population responds.\n\nToday, teams rely on surveys and historical data to help derisk those bets.\n\nBut those tools are inherently outdated and non-comprehensive: they observe people individually and at a point in time.\n\n_Real populations don’t behave that way._\n\nPeople influence one another. Reactions spread between groups. Adoption compounds or fades. Sentiment changes. A decision that resonates with one audience can change how another audience responds. Those interactions are what ultimately determine outcomes like revenue, conversion, adoption, and sentiment.\n\nCurrently, the only real “simulation” is the launch itself.\n\n## What we do\n\nStudio builds live models of a company’s market from real data on how consumers behave. It overlays internal data like sales and campaign performance with external evidence like social conversations and market trends.\n\nStudio uses those signals to construct highly specific audience cohorts and model how their behavior and influence interact over time.\n\nTeams can then:\n\n* **Test-run launches:** see how a new product, price, campaign, message, or initiative will unfold.\n* **Forecast outcomes:** revenue, conversion, adoption, sentiment, and other KPIs.\n* **Understand why:** know which groups drive the result, by how much, and trace outcomes to exact evidence.\n* **Test alternatives:** compare different versions of a decision to find the optimal variation.\n* **Improve over time:** compare forecasts with real outcomes and recalibrate the model.\n\nInstead of predicting “How will this person respond to X?”, Studio answers:\n\n“How will a population interact with X, and why?”\n\nStudio has live-tested with large consumer brands like Danone and Mila, and our model’s predictions match real-world KPI outcomes with >95% accuracy.\n\n## Why we built it\n\nWe’ve spent years predicting behavior. Nikita built simulation software to understand and predict individual decision-making at MIT, while Cameron built predictive models for financial institutions and in private equity, and agent-based society simulations at UC Berkeley MET.\n\nWe met at the same gap: you can simulate how a protein will fold or how a financial system will react to a market shock, but there is no equivalent for simulating how society will interact with a new idea. We’re now creating a studio to design society.\n\n## Where we’re starting\n\nWe’re initially working with enterprises making high-stakes audience-facing decisions across product, pricing, innovation, brand, marketing, and communications.\n\nBut the underlying problem is much broader.\n\nAny organization trying to create a particular response in a population should be able to test different versions of that decision, understand how reactions might spread, and compare likely outcomes before deploying it.\n\n## Our ask\n\nWe’d love to talk to teams making consumer-facing decisions. This could be a new product, pricing plan, campaign, etc.\n\nIf you have an upcoming decision you’d like to test before it goes live, we’d love to run it through Studio!\n\n[trystudio.ai](http://trystudio.ai)\\\n[team@trystudio.ai](mailto:team@trystudio.ai)\n\nLet’s chat: ", "url": "https://www.ycombinator.com/launches/Syy-studio-the-simulation-engine-for-market-behavior", "votes": 14, "created_at": "2026-08-21T18:00:00.302Z" } ] }, { "id": 32164, "name": "Prized", "slug": "prized", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/7162dc10af43dc1c38075eec773f7958383384d2.png", "website": "https://prized.dev", "all_locations": "San Francisco, CA, USA", "long_description": "Prized is a cloud development environment that feels local on your Mac. \r\n\r\nDevelopers get a dedicated, persistent machine with their repos, dependencies, and tools already configured. \r\n\r\nFiles sync into a real local folder, remote ports appear on localhost, and terminals persist through tmux, so developers and coding agents can keep working even after the laptop closes. \r\n\r\nMachines can be resized or cloned as workloads grow, without rebuilding the development environment.", "one_liner": "A cloud devbox that feels local", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1782800618, "tags": [ "Developer Tools", "SaaS" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/prized", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/prized.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/useprized", "x_url": "https://x.com/useprized", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/7162dc10af43dc1c38075eec773f7958383384d2.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 2447160, "full_name": "Hudson Griffith", "title": "Founder/CEO", "founder_bio": "Co-founder & CEO of Prized. Most recently he was an engineer at Suno where he shipped full-stack features to millions of users. Before that he was the first engineer at Gander (YC F24), a 0-to-1 fleet-management platform that was acquired by Archer Aviation (NYSE: ACHR), where he owned the core architecture and built AI voice and document agents. He studied Computer Science at the University of Florida.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/hudsongri/", "x_url": "https://x.com/hudson_gri", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9dc6b00d5a3a7c347781953481e598f0833b5efc.jpg" }, { "user_id": 2485896, "full_name": "Marinos Eliades", "title": "Founder", "founder_bio": "Marinos is the co-founder of Prized. He studied Mechanical Engineering at Stanford University. Most recently he was an engineer at Gander (F24), a 0-to-1 fleet-management platform for the private jet industry and was acquired by Archer Aviation (NYSE: ACHR). He previously served in the National Guard of Cyprus.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/meliades", "x_url": "https://x.com/marinoeliades", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/851436432e6a4e9a6c9f158342b60aa8313f9e6c.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112144, "title": "Prized: A cloud devbox that feels local", "tagline": "The fastest way to run coding agents in the cloud without giving up your local workflow.", "body": "Hey everyone 👋 We're Marinos and Hudson, co-founders of Prized.\n\n**TL;DR:** Prized gives developers a dedicated devbox in the cloud that still feels local. Your files stay synced to a real folder, ports show up on [localhost](http://localhost), and coding agents keep running when you close your laptop.\n\nLaunch video: \n\n**The problem**\n\nCoding agents can work for hours now. But only if you keep your laptop open.\n\nYou start a refactor, a migration, or a big build. Then you need to leave. Close the lid and the agent dies with it.\n\nYou can move everything to a VM, but now you're doing remote dev. Your files live somewhere else. Your ports need forwarding. Your terminal disappears when the connection drops.\n\nRemote development still feels remote.\n\nit shouldn't…\n\n**The solution**\n\nWith Prized, one command puts your project on a dedicated devbox without giving up local development:\n\n* **Real files on your Mac.** Prized syncs both ways into a normal folder, so everything keeps working the way it already does.\n\n\n* [**Localhost**](http://Localhost)** stays [localhost](http://localhost)**. Servers running on the box appear on your Mac automatically, including loopback-only services and hot reload.\n\n\n* **Sessions stay alive.** SSH lands in a persistent terminal session on the box, so disconnecting or closing your laptop doesn't stop the work.\n\n\n* **Compute grows with the task.** Resize the box without rebuilding your environment or moving your project again.\n\n\n* **Secrets stay out of the agent's hands.** For brokered secrets, the agent only sees a placeholder.\n\n**The team**\n\nWe worked together at Gander (F24) building fleet management software for private jets. After the company's acquisition by Archer Aviation (NYSE: ACHR), Marinos continued school at Stanford, and Hudson went on to be an engineer at Suno.\n\nWe built Prized for ourselves after getting tired of having to keep our laptops cracked open all the time. A raw VM kept the process alive but gave up the local workflow we actually wanted.\n\nCheck Prized out at [https://prized.dev](https://prized.dev/), or reach out at [founders@prized.dev](mailto:founders@prized.dev).", "url": "https://www.ycombinator.com/launches/TAm-prized-a-cloud-devbox-that-feels-local", "votes": 17, "created_at": "2026-08-25T17:00:00.790Z" }, { "id": 105647, "title": "Prized: Lovable for internal tools", "tagline": "The fastest way for non-engineers to ship secure internal tools.", "body": "Hey everyone 👋 We're Marinos and Hudson, co-founders of Prized.\n\n**TL;DR:** Prized lets anyone at your company build and ship secure internal tools with AI. Company data comes pre-connected and scoped to what each person should see, every access is audited, and finished tools deploy behind your company's sign-in.\n\n**Launch video:** \n\n**The problem**\n\nBuilding software has never been easier. And the people who need internal tools most, in ops, support, and finance, are done waiting on engineering.\n\nSo they build it themselves. They rip Claude Code or Cursor, pull straight from company data, and ship a working tool in an afternoon. No permissions. No audit trail. No one watching where the data goes.\n\nSecurity is assumed to be fine because the tool is \"just internal.\"\n\nit's not…\n\n**The solution**\n\nPeople aren't going to stop building with AI. So we built them a place where it's safe.\n\nWith Prized, anyone can spin up a sandbox that comes with company data already attached, scoped to what they should see, with an audit trail on every access. \n\nFrom there it's whatever building style they like: describe the tool in our app builder and watch it take shape, or bring a coding agent like Claude Code and build it themselves. Shipping is one click, and the finished tool sits behind your company's sign-in.\n\nThe person who runs the workflow finally gets to build the tool for it. Safely, this time.\n\n![uploaded image](/media/?type=post&id=105647&key=user_uploads/2485896/e426b427-8a79-4ca5-b462-b26b98db43e0)\n\n**Secure by default**\n\nSecurity is enforced in the infrastructure, not left to the person building the tool:\n\n* **Least privilege per tool.** Every tool gets its own database role and integration scopes; it can read and write only what it was granted.\n* **Credentials stay out of reach.** Secrets live in a broker and never touch the tool's code or the agent's context, so they can't leak or cross workspace boundaries.\n* **Outbound traffic proxy.** External calls only go to explicitly allowed hosts, and integrations are read-only until someone turns writes on.\n* **Admin visibility.** IT sees every tool in the org: who built it, what data it reads or writes, who can use it, and risky changes (like a destructive migration) need explicit approval before they ship.\n\n**Your agent, our sandbox**\n\nAlready have a favorite coding agent? Bring it. It plugs into the same governed sandbox and ships the same way.\n\n![uploaded image](/media/?type=post&id=105647&key=user_uploads/2485896/fcfa68c5-8c9f-4d13-b37c-d526e733539a)\n\n**The team**\n\nWe worked together at Gander (F24) building fleet management software for private jets. After the company's acquisition by Archer Aviation (NYSE: ACHR), Marinos continued school at Stanford, and Hudson went on to be an engineer at Suno.\n\n**Our ask**\n\nIf any of these describe you, we'd love to talk:\n\n* You work in ops, support, finance, or RevOps and keep waiting on engineering for small tools.\n* You lead engineering and one-off tool requests keep eating your roadmap.\n* You're in security or IT, wondering where AI-built tools are showing up at your company.\n\nCheck us out at[ https://prized.dev/](https://prized.dev/), or reach out at [founders@prized.dev](mailto:founders@prized.dev).\n\n![uploaded image](/media/?type=post&id=105647&key=user_uploads/2485896/223239da-19fa-44fb-9bd5-d18f5ecfdc74)\n\n", "url": "https://www.ycombinator.com/launches/RTz-prized-lovable-for-internal-tools", "votes": 10, "created_at": "2026-07-14T13:45:51.146Z" } ] }, { "id": 32167, "name": "Orchestra", "slug": "orchestra", "former_names": [ "Understudy Labs" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/900e0fabf138d248123ff3ddf28833736fbfb71d.png", "website": "https://orchestra.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Orchestra is an inference cloud. We capture your AI work traces, train smaller models, and automatically deploy them (only when they outperform the expensive model you are currently paying for).\r\n\r\nInstead of renting the same frontier model forever, your AI infrastructure gets faster, smarter, and cheaper every time it runs.\r\n\r\nSuppose your product uses Claude for a repetitive legal workflow.\r\n\r\nOrchestra observes successful production runs, learns the specific tools and judgment required for that workflow, and trains a specialized open-weight model.\r\n\r\nOnce that model matches or beats Claude on a held-out evaluation, Orchestra starts routing production traffic to it.\r\n\r\nAs more work is completed, the model keeps improving.\r\n\r\nWe started Orchestra because we believe every company using AI should be accumulating intelligence, not accumulating API bills.", "one_liner": "Self-optimizing neocloud that cuts LLM costs by 100x", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1783740567, "tags": [ "Open Source", "AI", "ML" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": true, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/orchestra", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/orchestra.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/understudy-labs", "x_url": "https://x.com/understudylabs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/900e0fabf138d248123ff3ddf28833736fbfb71d.png", "app_video_url": "https://bookface-images.s3.us-west-2.amazonaws.com/bf-public_profile/32225/app_video.mp4?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ASIAQC4NIECAOYE2QNZ5%2F20261020%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20261020T022739Z&X-Amz-Expires=60&X-Amz-Security-Token=IQoJb3JpZ2luX2VjEHEaCXVzLXdlc3QtMiJIMEYCIQDZr7mht6VpZroejzIM4ovdqqWjmh9u8AZ4wJ%2BcxZeHaQIhAN0TyP1C5dfPpTpdf3vcwz%2B044IBUbnJ55rbsOHYRWQWKuUDCDoQABoMMDA2MjAxODExMDcyIgxycpAO09Mq3A1cvyQqwgN39qg2Ox70YWArW8Z19bx1LbIV2O3f8r7xSc%2BDpl3XtmT02s9DHm%2FdyEMimOoBk5McFrR%2FThfgO9h1Zcmt1NqN1y6d1lWIsDAdluAWIqXM089dX1KVjFmsSwg8hWS2%2F4ysZFTLuD5VBUHeANJPyfmRKN9Kb4JepbPSE01eipdz4aqPxRqz%2FVUJEYLUTh5YJM6jJKNzl5l2mnWSSFiFdUfT5OS16LHDCYH4PGt9WRY2iseRxMu7FZcICWM0P8aN14WLxKsu1H0yVxk8LS%2BNbXezGhQnkMsa3bf7aMNU7x4TSs7KyM01QgiCZxC01YV43U4O8vFENPThr%2FHtKaAvtZ0%2BXSTwUE691O8mG15thNVkorInjIhilxdazy%2BzEYcGsAMNZtfx0Bd5Xn43Xf7b%2Fie4vv0d0RCHqbCGLmxRHj5VG3Pdta5Kju3%2Fpb8n%2FKmRjY%2FtuIfi9oTYAjc4PVyikHTm8UXURiHmaM0tqXZjxyh0iIponEKehSmtjmjudiXLWICFBoPhEXeaUeyK2AT6btvH4fclsQa0qRmkxdgguiWvrYxLKhNNfPLTc5QA0jlPSRoDJ%2F%2BxNHLFrGusOX7IwU6hfRMw6P3a1gY6pAETos2BWzWJY973oh8s1gy0zFzSYQP7N%2B39nIjIA7I0ymOVnRNJktoZRF6GV9DH3ymBPT2fnMgKvEaFAqVR%2F0WVwXz3ixgSwFGrvlwPr39cuFnKBEcdnivQJOXkexjqUxi9alVn76GV8ONTXD1ZupAQeLYI09gLaaAmJlyJPRAUCybD64XlECdPqS2TV2e%2FZlA5duJD1GekpDGthoc0y8PIc1ahwA%3D%3D&X-Amz-SignedHeaders=host&X-Amz-Signature=993a52d820dba9b6623af23273922c1e1980795db53cc2f2ee2ad22f445bb5e7", "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 3647059, "full_name": "Luis Manrique", "title": "Founder/CEO", "founder_bio": "Founder at Orchestra (YC S26). Previously founding Member of Technical Staff (and founding sales guy) at Gumloop, closing ~$2M in first year. Led Ads ML and also consumer AI as first PM at Instacart's Carrot AI. Prior co-founder/CEO at Ones and Zeros (text-to-SQL using transformers). Early at Wildfire (acquired by Google), then 5 years across Google's ads ML and programmatic platforms. SVP Solutions at VideoAmp building TV measurement infrastructure.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/luismanrique", "x_url": "https://x.com/lluismanrique", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/704197b652568953035fd4d4a70a73910a60a37c.jpg" }, { "user_id": 73303, "full_name": "Aamir Poonawalla", "title": "Founder/CTO", "founder_bio": "Founder at Orchestra (YC S26). 10 years at Instacart. Built the ads serving and ML infrastructure, auction platform, and experimentation framework. 2nd time YC Founder. MS in Computer Science from Georgia Tech.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aamir-p-a3575318/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4c9a96d07312de424f0f4d4023f16b6a3238ce54.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 114099, "title": "Orchestra: Self-optimizing inference cloud to cut your AI costs by 100x", "tagline": "Your AI inference gets smarter and cheaper the more you use it. ", "body": "We’re Aamir & Luis and we’re excited to launch our company’s new brand ahead of demo day this week: Orchestra ([orchestra.ai](http://orchestra.ai)) \\\n\\\nCompanies move their frontier inference spend to Orchestra, with a single URL swap. We relay their traffic through to the frontier labs, while building evals in the background automatically. These evals allows us to post-train specialist models that replace the frontier, at a fraction of the cost. When our model beats the frontier at your product’s work, we promote it and your bills go down. Then we repeat the loop continuously, forever; your AI inference gets smarter and cheaper the more you use it. \\\n\\\nWe’re going straight after the $104B inference market; Anthropic grew 14x YoY but we’ve seen time and time again that start-ups and enterprises are willing to move spend if you can prove the performance. That it. We don’t have FDEs. We don’t pre-train models. Most of the application layer lives in a CLI that coding agents use. \\\n\\\nWe just do one thing really well: building and routing your work to the most efficient model. We found the easiest way to build a fast growing business is to _sell_ your product to people who have pain and who are already spending money with someone else. The more we focused on the fastest growing segments of our business, the faster our product got better at solving a simple, but incredibly painful problem: “You don’t need a PhD to work in the mailroom”. \n\n**It’s the “Age of the Unc”** \n\nI shared my experiences doing YC as a 38-year old () but for any founders out there thinking about coming back to YC: do it. We worked 60-hour weeks, not 100-hour weeks. We had a network of angels, customers and advisors. We ran the marathon, instead of sprinting it. At 38, I’m old enough to understand that pedagogy is the point of YC. Learning how to work on your company is the point. You’ll get **so much more** out of YC when you’re doing it again with the hindsight of all your past successes and failures. Having done it before is a huge advantage, but having a long career in which I’ve encountered a lot of situations and had to make a lot of decisions and a few mistakes, and see others do the same, allows me to understand how YC’s advice impacts the real world. You will manage your agents better. You will close customers faster. You’ll build a better company. \n\n**Why Orchestra?** \n\nBranding is everything, especially outside of silicon valley. We launched as the ‘alternative’, the ‘backup’; but as time went on, we realized how open companies are to moving **all** their spend to a single place and making it work efficiently. Inference is the fastest growing market in the history of capitalism and we needed a brand that reflected our ambitions. Orchestra is going to win a huge chunk of that market; not the back-up, but the entire show. \n\nSay hello to Orchestra. And if your bills are too damn high, send us a note. We’d love to help.\n\n![uploaded image](/media/?type=post&id=114099&key=user_uploads/3647059/d32c3f74-5cd4-4026-ad40-1c7bd729a08a)\n\n", "url": "https://www.ycombinator.com/launches/TgJ-orchestra-self-optimizing-inference-cloud-to-cut-your-ai-costs-by-100x", "votes": 5, "created_at": "2026-09-07T14:00:00.839Z" }, { "id": 109219, "title": "Understudy: The self-optimizing inference cloud", "tagline": "We distill your production usage into cheaper, specialized models your company owns", "body": "Hey YC! We’re Luis and Aamir, founders of Understudy.\n\n### TL;DR\n\nWe help you cut your company’s Anthropic bill by 80% with no loss of performance.\n\nUnderstudy is an inference cloud. We capture your employee’s AI work traces, train smaller models, and automatically deploy them (only when they outperform the expensive model you are currently paying for).\n\nInstead of renting the same frontier model forever, your AI infrastructure gets cheaper and better every time it runs.\n\nSee more: \n\n### The problem\n\nLast week, we met with a company spending $40M/year on Anthropic. And it’s only going in one direction.\n\nTo them, it feels totally out of control.\n\nThey’re not alone - every company we meet is spending record amounts on models from frontier labs.\n\nBut almost none of that spending compounds.\n\nI.e. You send a prompt, pay for the answer, and throw away the experience. Six months later, you are still renting the same general-purpose model at roughly the same price.\n\nMeanwhile, training custom models requires an entire ML team to collect and clean traces, build evals, choose training data, fine-tune and reinforce models, test serving configurations, compare cost, latency, and quality, safely deploy or roll back the winner.\n\nFor most companies, it’s just too much to bother with.\n\n### What Understudy does\n\nPoint Understudy at an AI workload.\n\nWe capture how your existing agent performs, turn those traces into evals, and start training smaller models that can do the same job better.\n\nUnderstudy handles the entire loop:\n\n**Observe → evaluate → train → test → deploy → repeat**\n\nA new model is promoted only when it beats the incumbent on the metrics you care about, such as:\n\n* task success\n* inference cost\n* latency\n* privacy\n* deployment size\n\nYour customized models are owned by you, and can run in our cloud, your cloud, or on-premise.\n\n### An example\n\nSuppose your product uses Claude for a repetitive legal workflow.\n\nUnderstudy observes successful production runs, learns the specific tools and judgment required for that workflow, and trains a specialized open-weight model.\n\nOnce that model matches or beats Claude on a held-out evaluation, Understudy starts routing production traffic to it.\n\nAs more work is completed, the model keeps improving.\n\n### Why us\n\nAamir and I met building Instacart’s optimization and experimentation systems, growing multiple business lines to more than $1B in annual revenue.\n\nWe started Understudy because we believe every company using AI should be accumulating intelligence, not accumulating API bills.\n\n### Our ask\n\nSpending too much on repetitive model workloads? Ready to own your own models? We want to meet you - come to [understudylabs.com](http://understudylabs.com) today.", "url": "https://www.ycombinator.com/launches/SPb-understudy-the-self-optimizing-inference-cloud", "votes": 16, "created_at": "2026-08-05T02:21:26.322Z" } ] }, { "id": 32168, "name": "Synapse Semiconductor", "slug": "synapse-semiconductor", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/37650fbbc73ea1ea0b895fd819e93836825d1520.png", "website": "https://synapsesemi.org", "all_locations": "Durham, NC, USA", "long_description": "We want to push machines to where humans cant go. \r\nPresidential early career award winning Duke Semiconductor ECE Faculty Tania Roy and Sanjeev Chauhan are building a frontier semiconductor research lab. Their focus is on solving physical AI challenges from the semiconductor device level and building products around the moat of impossible to copy innovation. Their first work is on a new vision sensor that can run neural network filters right in the camera pixels and only collect data that machines care about. As physical AI develops, machines needs to understand the world around them if they aim to interact with them.\r\n", "one_liner": "New semiconductors that merge compute and vision", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1785109099, "tags": [ "Edge Computing Semiconductors", "Hard Tech", "Robotics", "Defense", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/synapse-semiconductor", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/synapse-semiconductor.json", "year_founded": null, "city": "Durham", "country": "US", "linkedin_url": "https://www.linkedin.com/company/synapse-semiconductor/", "x_url": "https://x.com/SynapseSemi", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/37650fbbc73ea1ea0b895fd819e93836825d1520.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 1383306, "full_name": "Sanjeev Chauhan", "title": "Founder/CEO", "founder_bio": "I like to break things and build things\n\n2 IEDM publications, 3 years Seed Stage VC, scaled deaplearning.com to 300k users, Stanford Linear Accelerator National Lab ML Researcher, IPAC 1st Author", "is_active": true, "linkedin_url": "https://linkedin.com/in/sanjeev-one", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ed6ba01db61837566ea4234d9d7ec34da6e0e4c1.jpg" }, { "user_id": 3682345, "full_name": "Tania Roy", "title": "Founder", "founder_bio": "Academic by profession, entrepreneur by nature. Love to control the user interface to reality.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/tania-roy-ba272b24/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/22e5f2656b361e2b88271fcd468b9cc08d0ba91f.jpg" } ], "jobs": [ { "id": 106484, "title": "Founding Engineer - RETINA", "url": "https://www.ycombinator.com/companies/synapse-semiconductor/jobs/T7HUC4J-founding-engineer-retina", "location": "San Francisco, CA, US / Durham, NC, US", "type": "Full-time", "role": "Engineering", "role_type": "Electrical", "salary_range": "$1M", "equity_range": "10.00%", "experience": "1+ years", "visa": "US citizen/visa only", "skills": [ "Verilog", "VHDL", "Computer Vision" ] } ], "news": [ { "title": "Duke ECE Hardware Startup Selected for Y Combinator | Duke Pratt School of Engineering", "url": "https://pratt.duke.edu/news/synapse-semiconductor-y-combinator/", "date": "Jun 08, 2026" } ], "launches": [ { "id": 107691, "title": "Synapse Semiconductor: We put a neural network inside the camera pixel.", "tagline": "Semiconductors to push machines where humans cant go.", "body": "\n\n**TLDR:** Synapse Semiconductor builds image sensors that process **AI on the sensor** itself. Each pixel captures light and runs neural-network computation in the same device, replacing the need for a GPU while cutting the power and latency of edge vision.\\\n\\\n\\\nA camera has no use by itself. Its value is only when its information gets evaluated. How can a camera that only has value when combined with a GPU support local intelligence? GPUs will get better, yet the fundamental problem is the separation of the camera and the GPU. Solving that separation will unlock bountiful innovation in machine perception. \n\n**RETINA: a chip designed like the human retina which runs neural network operations in the camera pixel.** \n\n\\\nWe have compressed the whole edge AI vision hardware stack into **one wafer**. CNNS, VLMS, etc can run on just one chip that senses light. Our compute transistors are also the same photosensors that see light. There is no longer separation between compute (Nvidia Jetson Nano) and a camera (RealSense Depth Camera). \\\n\\\nIf you are building robotic perception, drones, and vision systems for physical AI —> [book.synapsesemi.org](http://book.synapsesemi.org)\n\n![uploaded image](/media/?type=post&id=107691&key=user_uploads/1383306/bfa2012d-1c74-47b5-bdad-3127e94cd8bf)\n\n", "url": "https://www.ycombinator.com/launches/S0x-synapse-semiconductor-we-put-a-neural-network-inside-the-camera-pixel", "votes": 57, "created_at": "2026-07-27T03:40:58.979Z" } ] }, { "id": 32169, "name": "Speko", "slug": "speko", "former_names": [ "Speko Group, Inc." ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/a79212fefd832b4ac7d28da1ad3a7bc206dffaa1.png", "website": "https://speko.ai/?utm_source=ycombinator&utm_medium=profile", "all_locations": "San Francisco, CA, USA", "long_description": "Speko is a router for voice models: one API key, and your voice agent runs on the speech-to-text, LLM, and text-to-speech models we measured as best for your language and use case. The measurements are public at benchmarks.speko.ai: we test new models the day they ship and re-route when the numbers change, so you rewrite nothing. We ran this bake-off by hand for four years at previous company, shipping production voice for Fortune 500s.", "one_liner": "OpenRouter for Voice AI: one API, every call on the best speech model", "team_size": 4, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1783393099, "tags": [ "Developer Tools", "AI", "Conversational AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/speko", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/speko.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://linkedin.com/company/speko-ai", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/a79212fefd832b4ac7d28da1ad3a7bc206dffaa1.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 451241, "full_name": "Beknazar Abdikamalov", "title": "Founder/CEO", "founder_bio": "Founder and CEO of Speko (YC S26). Previously co-founder and CTO of Hupo (voice AI; $14M raised; backed by DST Global and Meta), where he spent four years shipping production voice AI for fortune 500s in 10+ languages. Earlier a software engineer at Amazon; MS in CS from UNIST, Korea.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/abdik/", "x_url": "https://x.com/beknabdik", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9a75e51d58a4fc108e1eb811218e0697d5c41ffe.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108267, "title": "Speko: OpenRouter for Voice", "tagline": "Every speech model, benchmarked language by language, wired into one API.", "body": "Speko is a router for voice models. One API key, and your agent runs on the speech-to-text, LLM and text-to-speech models we measured as best for your language and use case.\n\n![uploaded image](/media/?type=post&id=108267&key=user_uploads/451241/2038963f-c769-47a9-8fd9-c59f3966027b)\n\n### I picked voice stacks by trial and error for four years\n\nA voice agent is three models in a row: speech-to-text, then an LLM, then text-to-speech. Three vendors, a dozen credible options each. You run a bake-off, you argue about whose numbers are real, and you ship the combination that felt best that month.\n\nThen next week another model drops. And another. Nobody has time to re-run the bake-off, so the stack you picked under deadline quietly becomes the stack you are stuck with for a year.\n\nThat is how most voice agents are running today. Not on the best models. On the models that were best whenever someone last had a free week.\n\nPublic leaderboards do not settle it either. They are averages over somebody else's audio, in somebody else's conditions.\\\n\\\n[https://drive.google.com/file/d/1AjC8WVA96kUBdD09RJMUgFfvxGfxPvL-/](https://drive.google.com/file/d/1AjC8WVA96kUBdD09RJMUgFfvxGfxPvL-/view?usp=sharing)\n\n### We measure, then we route\n\nWe benchmark the speech models people actually ship on, language by language, and publish everything at [benchmarks.speko.ai](http://benchmarks.speko.ai): word error rate, finalize latency, time to first token, cost per minute. When a vendor ships a new model we test it the same day and the board updates.\n\nPoint your agent at Speko and that measurement becomes routing. One API key across every provider we support. Tell us the language and the conditions, and your traffic runs on the stack our numbers say wins. When the numbers change, the routing changes and you rewrite nothing.\n\nIf a provider fails when we open the connection, the router drops to the next ranked model on its own.\n\n### Or build the agent with us\n\nYou do not need an existing agent to start. A few lines gets a working voice agent live on Speko, and it inherits the same routing from day one.\n\nIf you are picking a voice stack right now, start with the numbers at [benchmarks.speko.ai](http://benchmarks.speko.ai), then build on them at [speko.ai](http://speko.ai).", "url": "https://www.ycombinator.com/launches/SAF-speko-openrouter-for-voice", "votes": 14, "created_at": "2026-07-29T19:10:13.923Z" } ] }, { "id": 32170, "name": "Gutgutgoose", "slug": "gutgutgoose", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/3b10bdb631232584efc1a8c09e67b74f8b953437.png", "website": "https://gutgutgoose.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Most data companies pay for their data. Ours pays for itself.\r\nCustomers send us a stool sample. We sequence it, build a metabolic model of that gut, and simulate whether each probiotic strain has an open niche it can colonize. We formulate the best candidates, then re-sequence 90 days later to see what actually took hold and what changed. Each cycle gives us the baseline gut, the intervention, and the outcome. At scale, the endgame is models that predict chronic disease years before symptoms.", "one_liner": "We sequence your gut and build the probiotics it’s missing", "team_size": 7, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1785190964, "tags": [ "Consumer", "Biotech", "Personalization", "Health & Wellness" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/gutgutgoose", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/gutgutgoose.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/gutgutgoose/", "x_url": "https://x.com/Gutgutgoose", "fb_url": "https://www.facebook.com/share/18ahzDMEav/?mibextid=wwXIfr", "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/3b10bdb631232584efc1a8c09e67b74f8b953437.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 991771, "full_name": "Leon Mojarrabi", "title": "Founder", "founder_bio": "started university at 14, spent four years at a yc tech unicorn, then launched a pharmacy that hit $4M in revenue to fund GutGutGoose; strain-level metabolic modelling of the gut to catch disease decades early. I'm unreasonably interested in your poop's data.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/leon-moj", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c69cb9d3ddd32f9a50e604a7cee8506f0be7a15d.jpg" }, { "user_id": 3677890, "full_name": "Anis Mihrshahi", "title": "Founder", "founder_bio": "Youngest Math Olympiad champion in state's history. \n\nAt GutGutGoose, I work operations behind the scenes and co-built the mechanistic algorithm behind our personalised probiotics.\n\nFun Facts: World Catan #1, Australian Powerlifting Champion, State Discus Champion, organised national conferences that brought 2000+ Australians in to think about creating meaningful change in society.\n\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/anis-mihrshahi-2a342a336/", "x_url": "https://x.com/anismihrshahi", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5ee2a444b14c5548c9ccc5cd1fe8814527d86097.jpg" } ], "jobs": [], "news": [ { "title": "Meet The YC Startups Betting On What Comes Next", "url": "https://www.forbes.com/sites/dariashunina/2026/09/03/meet-the-yc-startups-betting-on-what-comes-next/", "date": "Sep 03, 2026" } ], "launches": [ { "id": 108277, "title": "Gutgutgoose: Personalized probiotics", "tagline": "GutGutGoose builds personalized probiotics guaranteed to colonize, then uses every retest to train models of the human gut.", "body": "# TL;DR: We’ve created the last probiotic you’ll ever need - personalised, constantly improving, while also generating valuable data.\n\n### \n\n# What we do and why\n\nYour gut is the most unique part of your body, different enough that the same probiotic can work for one person and do nothing for another. Most probiotics disappoint because they’re one formula for everyone, and in any gut most of their strains never establish. GutGutGoose makes a probiotic that won’t work for any gut but yours: we sequence your microbiome, find which strains can establish in the microbes you already have, compound those into your formula, and re-sequence you weeks later to confirm they took hold.\n\nIt’s for people who’ve tried things for IBS, bloating, digestion, mood, weight or GLP-1 symptoms without luck. If it doesn’t work you get your money back, and retesting shows how your gut changed, so you know it’s not a placebo.\\\n\\\n\n\nThe innovation is accurately predicting which bacteria will live in a gut, which is extremely hard. It’s built on two peer-reviewed models, MICOM and AGORA2, and more than 7,000 metabolic reconstructions of gut bacteria, and it finds the strains that can hold in your gut’s community. On two published human cohorts, its predictions matched what actually established 84.7% and 76.7% of the time, and we reproduced the first at 78.1% ourselves. Off-the-shelf products manage under 5%. Once we can control which bacteria live where, we can run precision interventions in the gut and engineer it.\n\nWe’re a supplement today and can’t treat disease, but each cycle produces a record of one gut, one intervention, and the result. That record makes each customer’s next formula better, and across thousands of people it becomes the first longitudinal dataset of the human gut responding to known, controlled changes. Our vision is to use changes in metabolic pathways over time to predict Alzheimer’s, dementia and Parkinson’s a decade before symptoms appear.\n\n![uploaded image](/media/?type=post&id=108277&key=user_uploads/991771/50caffe3-e0c9-44e5-9212-da1bbabc0e00)\n\n# Traction\n\nWe gave a talk at an endometriosis event, and 20 of the 50 people joined our waitlist. The first night we opened paid deposits, we called five people from that list at 7:30pm. Three picked up, and we managed to get seven sign ups, they were so excited they referred and paid for their families that night. Five of the 20 referred someone else, and our traction has only grown since.\n\n# Team\n\n![uploaded image](/media/?type=post&id=108277&key=user_uploads/991771/13686614-1a12-4b14-a657-d33d042b3ee6)\n\nThe GutGutGoose founders winning the YC product pitch\n\nLeon started university at 14 and spent four years at a $3bn YC unicorn. After we started GutGutGoose, he built a pharmacy, now doing $4M a year, to fund the science behind our engine. Anis joined in September; he co-built the model, runs operations, and is the youngest ever to win Queensland’s Math Olympiad. He’s on a Chancellor’s Excellence Scholarship in AI and Data Science, and (fun fact) was once the world’s number one Catan player.\n\n# The ask\n\nWe’d love introductions to gut-health clinics, patient communities, or microbiome researchers who could help us recruit and shape the 80-person study. Email us at [admin@gutgutgoose.com](mailto:admin@gutgutgoose.com). More at [gutgutgoose.com](http://gutgutgoose.com).", "url": "https://www.ycombinator.com/launches/SAP-gutgutgoose-personalized-probiotics", "votes": 145, "created_at": "2026-07-30T15:49:21.242Z" } ] }, { "id": 32171, "name": "LemonLime", "slug": "lemonlime", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/8ffbde2e695962be35fad4bb6260c320a8d18271.png", "website": "https://www.lemonlime.com/", "all_locations": "San Francisco, CA, USA", "long_description": "LemonLime runs fully automated sales and marketing for small businesses.\r\n\r\nEvery morning, LemonLime finds you new leads and drafts personalized outreach, prepares branded content, and identifies untapped sales & marketing opportunities.\r\n\r\nAll you have to do is click send.", "one_liner": "Fully automated GTM for small business", "team_size": 5, "industry": "B2B", "subindustry": "B2B", "launched_at": 1783367890, "tags": [ "SaaS", "Sales", "Marketing", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/lemonlime", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/lemonlime.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/lemonlime-ai/", "x_url": "http://x.com/lemonlime", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/lemonlime", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/8ffbde2e695962be35fad4bb6260c320a8d18271.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3669391, "full_name": "Jordan Zietz", "title": "Founder/CEO", "founder_bio": "Run fully automated sales and marketing, no work required → lemonlime.com\nFormer CEO/Founder in social, scaling and distributing consumer apps.\nThe 44th Stanford Tree – yes, the one that got suspended.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jordan-zietz", "x_url": "https://x.com/jordna", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d6333e91c22d13f305beff908c0fdce4e240b39d.jpg" }, { "user_id": 3671976, "full_name": "Daniela Muñoz", "title": "Founder/CTO", "founder_bio": "Co-Founder & CTO of LemonLime, fully automated GTM for small business. \nPreviously co-founded Confetti, an AI-native social platform. Former SWE intern at Google and Microsoft. Dual BS in CS and HCI from Carnegie Mellon.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/danielamunoz12/", "x_url": "https://x.com/dani__munoz", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3116f1d883b3f5928512d22881999fc38d65f32e.jpg" } ], "jobs": [ { "id": 94768, "title": "AI Engineer", "url": "https://www.ycombinator.com/companies/lemonlime/jobs/odgsmPB-ai-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Backend", "salary_range": "$100K - $120K", "equity_range": "0.50%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [ "AI Agents", "Multimodal AI" ] }, { "id": 109541, "title": "Growth", "url": "https://www.ycombinator.com/companies/lemonlime/jobs/qeSCOOz-growth", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Marketing", "role_type": null, "salary_range": "$90K - $120K", "equity_range": "0.50%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 104861, "title": "LemonLime – Automate your workflows with a single click", "tagline": "AI that learns your business and self-creates automations, without writing a single line of code.", "body": "Hey everyone! We're [Jordan](https://www.linkedin.com/in/jordan-zietz/) and [Daniela](https://www.linkedin.com/in/danielamunoz12/), founders of LemonLime.\n\n**TL;DR**\n\nMost companies don't have the time or technical expertise to build custom AI automations. LemonLime connects to your existing tools, studies your business, and builds specialized automations for you — actual outcomes, fast.\n\n\n\n**The problem**\n\nEvery small business knows they should be \"using AI,\" but almost none of them can. We learned this firsthand doing custom AI consulting for businesses. After building workflows for a handful of customers, two things immediately became obvious:\n\n1. The context AI needs to succeed is scattered — files, documents, chat threads, you name it. Searching for and organizing this information was 95% of the work, every single time.\n2. Companies don't want to spend weeks learning another complicated new tool. That's a huge ask for busy teams.\n\nSmall businesses need impact out-of-the-box, they don’t have the time or capital to spend on fancy AI initiatives that aren’t creating value from day one.\n\n**Our solution**\n\nSince company information is all over the place, we started by building the layer underneath that powers AI search and retrieval. If you’ve heard the term “company brain”, that’s what this first layer is. \n\n![uploaded image](/media/?type=post&id=104861&key=user_uploads/3669391/f4cccbb6-1c98-4e2d-a37e-0d4603f8e3e5)\n\nThen, we took it a step further.  Value from day one.  \n\nAfter your company’s unique knowledge architecture is built, users can use plain-language to deploy agents and automations that support their business without writing a single line of code.  You tell it what you want automated, LemonLime automates it.  It’s that easy.\n\nDon’t know where to start?  No worries, LemonLime handles that, too. After running deep research on your business, LemonLime automatically surfaces suggested automations that you can implement with a single click.  \n\n![uploaded image](/media/?type=post&id=104861&key=user_uploads/3669391/cdca4a15-4d2b-4807-9014-51ad700dbeb0)\n\n\n**If you’re a small business, let’s talk <3**\n\nDon't have time to handle automating your sales, marketing, support, or something else? **We’ll set it up for you.** Email us at [hello@lemonlime.ai](mailto:hello@lemonlime.ai).\n\n\\\nOr, visit our website to get started: [lemonlime.ai](http://lemonlime.ai)\n\n![uploaded image](/media/?type=post&id=104861&key=user_uploads/3669391/84c1c154-5f99-4a95-82dc-d5031894ba8c)\n\n", "url": "https://www.ycombinator.com/launches/RHJ-lemonlime-automate-your-workflows-with-a-single-click", "votes": 12, "created_at": "2026-07-08T16:25:19.821Z" } ] }, { "id": 32172, "name": "Jcode", "slug": "jcode", "former_names": [ "Solo Labs" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/0d863d3cf6091c132bb3f8d22d57b3597bc971de.png", "website": "https://jcode.sh/", "all_locations": "San Francisco, CA, USA", "long_description": "Jcode is a high-performance coding-agent harness that makes it possible for developers to scale up multi-agent/swarm workflows. It's built from the ground up with a different architecture to solve the memory bottleneck (it's 20x more memory efficient than Claude Code) that has kept agent swarms from going mainstream. It has 20K GitHub stars, and throughputs $1mil per month of inference spend", "one_liner": "High performance coding agents", "team_size": 1, "industry": "Consumer", "subindustry": "Consumer", "launched_at": 1786012852, "tags": [ "Developer Tools", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Consumer" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/jcode", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/jcode.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/in/jeremy1huang/", "x_url": "https://x.com/1jehuang", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/0d863d3cf6091c132bb3f8d22d57b3597bc971de.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 2670902, "full_name": "Jeremy Huang", "title": "Founder/CEO", "founder_bio": "hacked github\ntop 500 monkey type\nhttps://github.com/1jehuang\nage 21", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jeremy1huang/", "x_url": "https://x.com/Jeremy069626792", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9ab93ad5c963e25acbd5597cd4094a99163c0c0b.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109551, "title": "Jcode: High-performance coding agents", "tagline": "20x more agents, 20x more productivity, 20x less RAM", "body": "[ ](https://drive.google.com/file/d/1zthUpXW7Hi4L-RS7sdzT1gbcEoS-VlM5/view?usp=sharing)\n\\\nThis is me running 20 agents in parallel. This is my real everyday workflow, not a demo.\n\n**What we do** Jcode is an open-source coding agent that's 20x more memory-efficient than Claude Code, which means you can run 20x more agents in parallel on the same machine. It hits state-of-the-art results on coding benchmarks. Try it: \n\n**The problem** Most coding agents are heavy. Each instance eats gigabytes of RAM, and because memory is the bottleneck, you can only run a handful at once. That caps how much work you can hand off at a time. The real power of agents is dispatching lots of different tasks in parallel and having them all worked on at once, and that's exactly what today's tools make too expensive to do.\n\n**Our solution** Jcode is engineered from the ground up to be lightweight, roughly 20x more memory-efficient than Claude Code. Since RAM is what caps how many agents you can run, cutting memory 20x lets you fire off 20x more tasks in parallel on the same hardware. It's fully open source, so you can run it yourself and inspect exactly what it does.\n\n**Who I am** I'm Jeremy, 21, building Jcode. I think running many agents in parallel is the single biggest lever in developer productivity. Previously I hacked GitHub Copilot for unlimited tokens, and I'm top 500 in the world on Monkeytype. GitHub: \n\n**Developers:** Try Jcode today: \n\n**Investors:** I'm raising around Demo Day. Reach out: [jeremy@solosystems.dev](mailto:jeremy@solosystems.dev)\n\n![uploaded image](/media/?type=post&id=109551&key=user_uploads/2670902/ee164e14-5c6a-4811-9e21-c0a80aa3c004)\n\n", "url": "https://www.ycombinator.com/launches/SUx-jcode-high-performance-coding-agents", "votes": 35, "created_at": "2026-08-10T16:49:30.145Z" } ] }, { "id": 32187, "name": "Insurf", "slug": "insurf", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/203aaca34875f8d2fd195e6c042414266aefc1ca.png", "website": "https://www.insurf.io", "all_locations": "Providence, RI, USA", "long_description": "Insurf is the AI-native data and decision layer for health insurance. Our model Surely identifies the lowest-cost coverage for anyone, saving individuals, families, and companies thousands per employee. Our product Inveto collects the data from hospitals and clinics that informs Surely.", "one_liner": "The AI-Native Decision Layer for Health Insurance", "team_size": 3, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1782286226, "tags": [ "Health Tech", "Healthcare", "Insurance", "Automation", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/insurf", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/insurf.json", "year_founded": 2026, "city": "Providence", "country": "US", "linkedin_url": "https://www.linkedin.com/company/insurf/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/203aaca34875f8d2fd195e6c042414266aefc1ca.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 3378177, "full_name": "Bryan Chung", "title": "Co-Founder", "founder_bio": "Co-founder at Insurf. Samsung at 14. 18th Internationally in Physics, Top 80 Internationally in Computer Science, 2x IEEE Publisher, USACO Platinum, Research at Cambridge, Brown, CMU, RIT, and UCSC", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/bryan-chung-21b531265/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6728ca29b0a22eb5a13e927f97bb5b9c2a9ef96e.jpg" }, { "user_id": 3378885, "full_name": "Chang Lu", "title": "Co-Founder", "founder_bio": "Co-founder at Insurf. Ex Brown BS/MD\nPreviously studied Neuroscience and Economics at Brown. Conducted zero-shot modeling research at Emory Winship Cancer Institute.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/changxianlu", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/bbd81a7330818b1c1cb7ebd17246c72911cb0146.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111220, "title": "Insurf: We save people and employers thousands on health insurance costs", "tagline": "We built a computation layer for health insurance that predicts the lowest true-cost coverage for anyone. We will be the national standard for how America chooses coverage, with the only engine that knows both what plans cost and what they actually do.", "body": "Hey everyone — we're **Chang and Bryan**, the founders of **Insurf**.\n\nAmerican health insurance is f\\*\\*\\*ed: chosen by guesswork, enforced by denial. We built the engine that fixes both halves.\n\n**TL;DR**\n\n**Surely** computes the lowest-cost plan for anyone. **Inveto** fights denials for real clinics across Texas, Georgia, and California, and each contract-authorized, de-identified outcome becomes coverage data nobody else has. Everything Inveto learns, Surely knows.\\\n\\\n[https://drive.google.com/file/d/1x5B44OPfZxd-32UBeGT-0a5WVdMf3Jnl](https://drive.google.com/file/d/1x5B44OPfZxd-32UBeGT-0a5WVdMf3Jnl/view?usp=sharing)\n\n**The problem: America picks health plans from one number**\n\nMore than 20 million Marketplace enrollees choose coverage off the sticker premium. That guess is expensive.\n\n• In our data, the lowest-premium plan was the lowest-cost option in **0 of 2,000 simulated years**.\n\n• In one example, the cheaper sticker would have cost **$8,052 more** in a single bad year.\n\n• That's one person. Multiply the risk across a family. Across a company.\n\n• This year alone, **four million Americans** found out what their plans actually cost — and walked away from coverage.\n\nWe compute.\n\n**Surely — the cost engine**\n\nSurely ranks all plans by what they'll actually cost: premiums, deductibles, drug tiers, networks, out-of-pocket caps. Every number sourced from government filings.\n\n**Inveto — the denial workbench**\n\nClinics upload a denial and the clinical record; Inveto produces source-cited, physician-signed appeals and supported prior authorizations. Live with real clinics across Texas, Georgia, and California today.\n\n**The moat is the join**\n\nEvery contract-authorized Inveto outcome becomes rights-cleared data about what plans actually do, including approvals, denials, appeals; Surely reads all of it. **What plans cost + what plans do. Nobody else has both.**\n\n**Why now**\n\nIn 2028, health plans stop renewing automatically: 20 million+ people will re-choose coverage every single year — the largest forced re-shop in the history of American health insurance. $150 billion a year in premiums, re-decided annually.\n\nAnd employers are next. The federal government projects **800,000 companies and 11 million+ people** will switch to ICHRA, where employees choose their own plans. That's a **trillion-dollar shift**. With Insurf, employers can expect to save **\\~$5,000 per employee, per year**.\n\n**Who we are**\n\nTwo friends who met during orientation at Brown. Chang got into medical school at 17, then worked a pharmacy counter watching thousands of people pay out of pocket for what they were promised. Bryan placed internationally in CS and physics competitions and has published IEEE research since high school.\n\n**Our ask**\n\n• Tired of overpaying for health insurance?\n\n• Weighing ICHRA, or trying to cut your company's benefits spend?\n\n• Running a specialty clinic buried in denials and prior auths?\n\nSee how much your company can save here: [insurf.io/audit](https://www.insurf.io/audit) or email [founders@insurf.io](mailto:founders@insurf.io) — we'll show you exactly what you'd save.\n\n**Everyone sells plans. Nobody prices them. We compute.**", "url": "https://www.ycombinator.com/launches/Svs-insurf-we-save-people-and-employers-thousands-on-health-insurance-costs", "votes": 97, "created_at": "2026-08-19T21:00:00.658Z" } ] }, { "id": 32188, "name": "IMPACT Drones", "slug": "impact-drones", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c7df3a5c9dfd6e7d9865e24f95f172c997a7183b.png", "website": "https://www.impact-drones.com", "all_locations": "San Francisco, CA, USA", "long_description": "IMPACT builds autonomous interceptor drones that ship in sealed containers. The box sits at the site it protects, data centers, power grids, airports, bases, with no crew on the pad. When sensors cue a hostile drone, an interceptor pierces its own frangible lid, launches, and rams the target, guided by onboard vision that keeps working when GPS is jammed. Legacy effectors cost $1M+ per shot against $30k threats; ours cost less than the threat and scale with the swarm, the same class of interceptor already downing Shahed drones over Ukraine. Sites subscribe to protection instead of buying missiles, hardware, reloads and monitoring included: Air Defense as a Service. 180M+ USD in signed letters of intent.", "one_liner": "Air Defense as a Service for Civilian Infrastructure.", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Defense", "launched_at": 1785702345, "tags": [ "Hardware", "Drones", "Swarm Robotics", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Defense" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/impact-drones", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/impact-drones.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/impactdrones", "x_url": "https://x.com/impactdrones", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c7df3a5c9dfd6e7d9865e24f95f172c997a7183b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 3245067, "full_name": "Jasper Lortije", "title": "Founder/CEO", "founder_bio": "Co-founder & CEO, Impact Drones. Aerospace engineer (MSc, TU Delft), over a decade building drones. Built the VTOL drone that won the Hannover Messe VTOL competition. Worked alongside special forces on the counter-drone security plan for the 2025 NATO Summit, protecting heads of state including President Trump.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jasper-l/", "x_url": "https://x.com/JasperLort", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a658fea832c15f888808bdf1131fe0fc62b3df85.jpg" }, { "user_id": 3677968, "full_name": "Alessandro Fiume", "title": "Founder", "founder_bio": "Co-founder at Impact Drones. Columbia MSc. Built the pipeline behind 180M+ USD in signed LOIs. At Bloomberg, drove 80% week-over-week growth in terminal activations closing 2M USD in contracts. Scholarship fellow to Columbia & ESCP, valedictorian of CIV, France.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/alessandro-fiume45/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/afa1db0b96ca55b32dea0b32e29b6a0fe999eda2.jpg" } ], "jobs": [ { "id": 114195, "title": "Founding Engineer, Robotics and Flight Software", "url": "https://www.ycombinator.com/companies/impact-drones/jobs/dpDZ9V6-founding-engineer-robotics-and-flight-software", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Robotics", "salary_range": "$120K - $170K", "equity_range": "0.50% - 1.25%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [ "C++" ] }, { "id": 114196, "title": "Integration Engineer, Radar and Drone Software (Counter-UAS)", "url": "https://www.ycombinator.com/companies/impact-drones/jobs/LKeth6O-integration-engineer-radar-and-drone-software-counter-uas", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Embedded systems", "salary_range": "$100K - $150K", "equity_range": "0.20% - 0.40%", "experience": "1+ years", "visa": "US citizen/visa only", "skills": [ "C++", "Python", "Linux", "Sensor integration" ] }, { "id": 114197, "title": "Autonomy Engineer, Interceptor Drones", "url": "https://www.ycombinator.com/companies/impact-drones/jobs/yVR7VxO-autonomy-engineer-interceptor-drones", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Robotics", "salary_range": "$110K - $160K", "equity_range": "0.25% - 0.45%", "experience": "1+ years", "visa": "US citizen/visa only", "skills": [ "C++", "Python", "Computer Vision", "Embedded Linux" ] } ], "news": [ { "title": "Impact launches Air Defense as a Service: autonomous interceptor drones in a sealed box, deployed at the sites drones are already hitting - EDR Magazine", "url": "https://www.edrmagazine.eu/impact-launches-air-defense-as-a-service-autonomous-interceptor-drones-in-a-sealed-box-deployed-at-the-sites-drones-are-already-hitting", "date": "Aug 17, 2026" }, { "title": "Y Combinator-backed Impact Drones to offer missile-intercepting drone service for data centers and grid operators - DCD", "url": "https://www.datacenterdynamics.com/en/news/y-combinator-backed-impact-drones-launches-offers-missile-intercepting-drone-service-for-data-centers-and-grid-operators/", "date": "Aug 17, 2026" }, { "title": "Impact Launches Air Defense As Service Using Drone Interceptors | Aviation Week", "url": "https://aviationweek.com/aerospace/advanced-air-mobility/impact-launches-air-defense-service-using-drone-interceptors", "date": "Aug 17, 2026" }, { "title": "Impact Drones launches autonomous “air defence as a service” for critical infrastructure - Unmanned airspace", "url": "https://www.unmannedairspace.info/counter-uas-systems-and-policies/impact-drones-launches-autonomous-air-defence-as-a-service-for-critical-infrastructure/", "date": "Aug 06, 2026" } ], "launches": [ { "id": 108922, "title": "Impact Drones: Air Defense as a Service for Civilian Infrastructure", "tagline": "Containerized drone interception for airports, ports, data centers and power plants.", "body": "# \n\nWe build containerized cells of autonomous interceptors that protect critical civilian infrastructure from hostile drones. Park the cell next to anything you need to keep running: it detects an incoming threat, launches and physically neutralizes the incoming aerial threat. [www.impact-drones.com](http://www.impact-drones.com)\n\n**OUR LAUNCH VIDEO**\n\n\n\n**TL;DR:** Half this batch is building AI agents. Your agents live in data centers. And these sites are now under threat. In March, drones took three AWS data centers offline in the Middle East and a dozen core cloud services with them. **Amazon lost 200 million USD in one day.** We sell a box you put next to anything you'd rather keep: an interceptor pierces its own lid, and the threat stops existing. We'll keep your agents running.\n\n**Why now:** The drone threat went industrial and air defense math didn't.\n\n**A single patriot missile costs more than 1m USD to shoot down a drone worth 30k USD.** Critical infrastructure needs a defense system that can be deployed at scale without losing the cost-per-interception battle.\n\n**What we're building:** a cell holding autonomous interceptors, always charged, always ready for when a threat arises. Unlike jamming, our system physically intercepts the threat. Unlike conventional air-defense systems, it is designed specifically for protecting civilian infrastructure affordably and at scale.\n\n**Why us:** two founders who ship hardware fast.\n\n[@Jasper Lortije](https://bookface.ycombinator.com/user/3245067): aerospace engineer (MSc at TU Delft), previously founded a Dutch company building GNSS-denied visual navigation for autonomous systems.\n\n[@Alessandro Fiume](https://bookface.ycombinator.com/user/3677968): sales and operations, ex Bloomberg, MSBA at Columbia.\n\n**Our ask:** intros to the people who own the airspace problem: data center security and infrastructure leads (hyperscalers and neoclouds especially, your uptime now has an air defense dependency), airport operations directors, and energy grid security teams. And if you're building agents: we're literally your uptime insurance, come say hi. [jasper@impact-drones.com](mailto:jasper@impact-drones.com) & [alessandro@impact-drones.com](mailto:alessandro@impact-drones.com)", "url": "https://www.ycombinator.com/launches/SKo-impact-drones-air-defense-as-a-service-for-civilian-infrastructure", "votes": 21, "created_at": "2026-08-04T18:02:22.616Z" } ] }, { "id": 32189, "name": "Verdant", "slug": "verdant", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/4de79d62c0767417bcc285f4339875d51c92131b.png", "website": "https://www.verdantapp.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Verdant sells AI agents to local government to make them more efficient. We're starting with permitting, the highest-volume workflow in local government that requires judgment. \r\n\r\nMany startups fail to sell to government, but not us - we've processed 200+ applications and gone from 0 to 4 paid deployments across 4 states with an average close time of 4 weeks.", "one_liner": "AI Agents to Make Local Gov Efficient", "team_size": 2, "industry": "Government", "subindustry": "Government", "launched_at": 1782774541, "tags": [ "GovTech", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Government" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/verdant", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/verdant.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/verdantapp", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/4de79d62c0767417bcc285f4339875d51c92131b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/ba7cd28e387491131be693a84de0bb0db99c7a71.png" ], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3242750, "full_name": "Aidan Ng", "title": "Founder", "founder_bio": "Co-founder & CEO of Verdant. Before Verdant, I built commercial real estate underwriting models as a business analyst at Capital One. \n\nBefore that, I studied International Politics at Georgetown and spent three years in the government space at DOJ, USDA, the Hill, and Deloitte Government and Public Services.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aidan-c-ng", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/06f5b92036568e3a0f58f53ff7de3ee788672855.jpg" }, { "user_id": 1527348, "full_name": "Jason Yi", "title": "Founder", "founder_bio": "Co-founder of Verdant. Before Verdant, I built AI HR software for local government at Holly and legislative tracking tools at Aidy (YC W24).\n\nPreviously, I studied computer science at Georgetown where I researched automation for public benefits delivery, AI safety, and published data security research.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jason-yi1/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5e581a5ee73cd1068b17f8bc9cabb05fd43966b8.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 113860, "title": "Data center development is broken. We fix that.", "tagline": "Helping American local governments steward American industrial innovation", "body": "Local governments are the backbone of industrial development - through zoning codes, they determine where, when, and how anything gets built.\n\nToday, data centers are landing in cities whose zoning codes never imagined them, and local governments are scrambling to respond. They need months of work before a single line of new code gets drafted - that means comparing their code with other cities, checking for consistency with state law, and more.\n\n![uploaded image](/media/?type=post&id=113860&key=user_uploads/1527348/6d052667-b707-4790-a425-c4173d7010dd)\n\nWe built Verdant Code to compress months into days. We automate the sourcing and the analysis so cities spend their time on judgment instead of data retrieval. Long-range planning work that used to take a six-figure consultant engagement now happens in-house as cities intended.\n\n![uploaded image](/media/?type=post&id=113860&key=user_uploads/1527348/cd930e43-f7cd-4c12-ade6-8d6d330d081f)\n\nData center development is a symptom of a larger problem with America's industrial base. Local governments run the code and permitting decisions that gate industrial development, but they run them on decades-old workflows. Local governments spend $11B a year on this work, yet development stalls anyway.\n\nAmerican zoning codes have not caught up with American innovation. It's time for that to change.", "url": "https://www.ycombinator.com/launches/TcS-data-center-development-is-broken-we-fix-that", "votes": 5, "created_at": "2026-09-04T19:00:00.099Z" }, { "id": 104860, "title": "Verdant: helping cities process permits faster", "tagline": "Modern permitting and planning software for local governments", "body": "**TL;DR:** \n\nVerdant is building the modern tech infrastructure for local government, and we’re starting with end-to-end permitting software that dramatically cuts local government permit review time. \n\nWe help applicants file complete, compliant applications upfront, then help staff review them - our AI reads documents, plans, and diagrams, then flags issues and drafts the final report. \n\n**Launch Video**\n[ ](https://drive.google.com/file/d/1e5VeQc-NIrlNqIHZP-ASPhuE9DwUzouM/view?usp=sharing)\n**Our story**\n\nWhile studying international politics at Georgetown, I spent three years in and around government (Capitol Hill, then DOJ and Deloitte) and kept hitting the same wall: policy sets direction, but execution is throttled by legacy software. My co-founder Jason has worked in govtech for the past four years researching public benefits automation at Georgetown and building govtech products at Holly and Aidy (YC W24) for local governments.\n\nJason and I connected over our passion for improving local government processes. The people who keep cities running are stretched thin across understaffed offices, and we want to give them the capacity to do more of what they signed up for. After interviewing 200+ local government staff, we learned that permitting is one of their most painful processes and their software has not done enough to help.\n\n![uploaded image](/media/?type=post&id=104860&key=user_uploads/3242750/71ea38cf-11ee-427f-8a30-f9381d983c70)\n\n**What we’ve heard**\n\nPermit and entitlement review takes too long, and local government staff are overwhelmed. Applicants often apply with incorrect information, triggering rounds of back-and-forth with the government. And once complete materials are finally in, understaffed departments can’t review everything quickly enough. \n\nPandemic-era incumbent contracts are churning because staff are frustrated with expensive legacy systems that only intake and track permits – no automation involved. Our end-to-end solution picks up where those systems stop, automating the review itself to fix permitting where it's most broken.\n\n**What we’re building**\n\nSay a homeowner wants to renovate their house. Before anything else, they have to figure out which permits and entitlements they need. Verdant starts here: applications chat with an AI agent trained on the city ordinances to determine what to apply for. From there, we route applications to a Turbo Tax-style permitting intake portal that guides them through the application, answers their questions, and flags gaps before submission, so what lands on the staff's desk is correct and compliant.\n\nAfter submission, our AI helps building officials and planners run completeness checks on applications, flagging errors in site plans and drawings. When the application is approved, Verdant uses the context already gathered to draft the next deliverable, whether that's a staff report for the planning commission or a memo for city council.\n\nBeyond the permit lifecycle, Verdant also supports long-range planning: our ordinance research tools let local governments compare how other jurisdictions have written their codes so they have reference points when updating their own.\n\n**Our asks**\n\n* Reach out! If you work in local government (planners, building officials, community developers, city managers, etc.) we’d love to chat.\n* Spread the word! If you know anyone in the permitting or planning space, we’d love to connect with them.\n\n \n\n**Contact us at [founders@verdantapp.com](mailto:founders@verdantapp.com)**", "url": "https://www.ycombinator.com/launches/RHI-verdant-helping-cities-process-permits-faster", "votes": 4, "created_at": "2026-07-07T16:56:24.689Z" } ] }, { "id": 32190, "name": "General Context Labs", "slug": "general-context-labs", "former_names": [ "General Context Labs", "Instaplay" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/0321fc7ea5809962e3910e985c4d9d2acbbed27a.png", "website": "https://generalcontextlabs.com", "all_locations": "San Francisco, CA, USA", "long_description": "General Context Labs builds products and datasets that push the limits of frontier models. Our initial breakout product, Instaplay is an interactive app-generation platform home to over 100,000 users who create and discover content.", "one_liner": "Building Products and Datasets for Frontier Models", "team_size": 4, "industry": "Consumer", "subindustry": "Consumer -> Gaming", "launched_at": 1784188651, "tags": [ "Gaming", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Consumer", "Gaming" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/general-context-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/general-context-labs.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/generalcontextlabs/", "x_url": "https://x.com/generalcontext_", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/0321fc7ea5809962e3910e985c4d9d2acbbed27a.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 957153, "full_name": "Akhil Aggarwal", "title": "Founder/CEO", "founder_bio": "Prev Investor at Bain Capital Ventures", "is_active": true, "linkedin_url": "https://linkedin.com/in/akhil--aggarwal", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fbf5746b7c5c33e0139756c7d4a6f3324688d392.jpg" }, { "user_id": 2530050, "full_name": "Frank Liu", "title": "Founder/CTO", "founder_bio": "Prev SWE at Instagram", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/frvnk/", "x_url": "https://x.com/frvnkliu", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3920d2baf1a54b1af10c29ed65fda6aa5f141e61.jpg" }, { "user_id": 1071066, "full_name": "Gary Wu", "title": "Cofounder", "founder_bio": "Prev APM at Salesforce", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/garywu23/", "x_url": "https://x.com/gary_wu23", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/22cab24670904f03e22bd711e013a7b42c882690.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109236, "title": "Instaplay – Make and play anything", "tagline": "Every game imaginable is about to exist, and you'll find it or make it on Instaplay. Make and play anything in minutes, then share it with friends.", "body": "**TL;DR:** The world is burning trillions on AI to make us more productive, but almost none of it has made our lives more fun. We're changing that. On Instaplay, anyone can describe an idea and play it with friends minutes later. 20,000+ creators and 100,000+ monthly players are already building: multiplayer zombie FPS campaigns, interactive classroom lessons, someone even proposed with a game.\\\n\\\n\n\n**The problem**\n\nThe last decade of social platforms optimized for content you consume by yourself. But the most durable thing online has always been doing things together, and games are the most natural way people connect. We think the next great social platform is one where friends play, not scroll.\n\n**What we built**\n\nDescribe a game, make it, and play it.\n\n* Solo and multiplayer\n* 2D and 3D, playable on the web and on Android (iOS coming soon)\n* Everything is social, with room codes, leaderboards, likes, comments, and creator profiles\n\nEvery game is a link, so it spreads through group chats and Discord instead of dying after one play.\n\n**Traction**\n\n20,000+ creators and 100,000+ monthly active users, growing organically, much of it through discovery via AI tools like ChatGPT. The loop is working: users make things, share them, and their friends come back to play and make their own.\n\n**Where this goes**\n\nGaming is already bigger than film and music combined, and Roblox showed that handing creation tools to players can build a $40B company. AI takes that further: when making a game costs one sentence, games stop being products and become a language, made in minutes, shared like memes, personalized to the exact people playing them.\n\n**Team**\n\n* **Gary Wu:** Prev APM at Salesforce, Harvard CS and US Presidential Scholar.\n* **Frank Liu:** Prev SWE at Instagram. Top-1,000 TFT player.\n* **Akhil Aggarwal:** Prev investor at Bain Capital Ventures, Engineer at AWS building credential-threat detection deployed globally.\n\nWe're lifelong gamers on a mission to make AI the most powerful medium for fun. The biggest thing AI changes about games isn't how they're made, it's who gets to play.\n\n**Our ask:** Try it and make something today at \n\nWe'd love intros for consumer and influencer partnerships. Reach us at [founders@instaplay.ai](mailto:founders@instaplay.ai)", "url": "https://www.ycombinator.com/launches/SPs-instaplay-make-and-play-anything", "votes": 14, "created_at": "2026-08-05T16:25:59.432Z" } ] }, { "id": 32192, "name": "Rise Reforming", "slug": "rise-reforming", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/5b908857ac82ef0d2a3b793f89fa198298587f8a.png", "website": "https://rise-reforming.com/home", "all_locations": "Chicago, IL, USA", "long_description": "Geopolitical conflicts and natural disasters frequently expose the vulnerability of the American chemical industry. The reason? Production is centralized and fossil fuel-dependent. The solution? Localized production using an undervalued resource: biogas. The US produces enough biogas to make over $20 bn/y worth of chemicals, but 60% is wasted in low-margin applications or flared altogether.\r\n\r\nRise Reforming is commercializing a modular process to transform this underutilized biogas on-site into cheap, supply-secure chemicals. Our beachhead market is dimethyl ether (DME), used as an aerosol propellant and a low-carbon fuel additive in propane. Our target market is methanol: an established billion $ chemical and the future of low-carbon shipping.\r\n\r\nWe're doing our pilot project at a Chicagoland wastewater plant. This will be the first wastewater biogas-to-methanol demo in the Americas.\r\n\r\nWe're hiring for a senior chemical process engineer. Apply here: https://www.ycombinator.com/companies/rise-reforming/jobs/wJ9Q9nv-senior-chemical-process-engineer", "one_liner": "We turn waste gases into supply-secure chemicals ", "team_size": 4, "industry": "Industrials", "subindustry": "Industrials -> Climate", "launched_at": 1783464355, "tags": [ "Hard Tech", "Smart Waste Management", "Alternative Fuels", "Renewable Energy", "ClimateTech" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Climate" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/rise-reforming", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/rise-reforming.json", "year_founded": 2024, "city": "Chicago", "country": "US", "linkedin_url": "https://www.linkedin.com/company/risereforming", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/5b908857ac82ef0d2a3b793f89fa198298587f8a.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3646927, "full_name": "George Rose", "title": "Founder/CEO", "founder_bio": "Co-founder and CEO at Rise Reforming (S26). Previously @ Nexamp, Highland Electric Fleets, Mantel Capture, GenH. Background in Molecular Engineering @ UChicago.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/george-rose25/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ee413aabedd5216b24ad28029e4413061ab156fa.jpg" }, { "user_id": 3620299, "full_name": "Lucas Zubillaga", "title": "Founder/CTO", "founder_bio": "Co-Founder & CTO at Rise Reforming (S26). Double B.Sc. in Molecular Engineering and Physics from the University of Chicago. Prev: nuclear waste management at Argonne National Laboratory, offshore wind at Avangrid/Iberdrola.\n\nLed the design, procurement, construction, and operation of a bench-scale reforming unit with controls. \n\nAlso have a Spotify playlist with 200k+ saves. Check it out at: https://open.spotify.com/playlist/3rV2gyEKHxixpKa3IDQLvF?si=3458c33e03e84738&nd=1&dlsi=d0006826be884d", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/lucaszubillagamaharg/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/209b2a1f6985cd73931fc7cb16aa2a3bb8462a42.jpg" }, { "user_id": 3610381, "full_name": "Jona van Oord", "title": "Founder/COO", "founder_bio": "Environmentalist turned engineer. Plan on spending the rest of my life building hardware to address the climate disaster :)\n\nCo-Founder and COO @ Rise Reforming (S26), Molecular Engineering @ UChicago", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jonavanoord/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e79cd2c8431a00a358ef4d9b12c2ee2e95e79849.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 105232, "title": "Rise Reforming: distributed chemical production", "tagline": "We turn waste gases into valuable supply-secure chemicals.", "body": "Geopolitical conflicts and natural disasters frequently expose the vulnerability of the American chemical industry. The reason? Production is centralized and fossil fuel-dependent. The solution? Localized production using an undervalued resource: biogas. The US produces enough biogas to make over $20 bn/y worth of chemicals, but 60% is wasted in low-margin applications or flared altogether.\n\nRise Reforming is commercializing a modular process to transform this underutilized biogas into valuable supply-secure chemicals like dimethyl ether (DME) and methanol. Our technology can be deployed on-site, helping biogas producers properly valorize their gas.\n\nWe've just broken ground on our pilot project at a Chicagoland wastewater plant. This will be the first wastewater biogas-to-methanol demo in the Americas.\n\nRise Reforming is hiring for a senior chemical process engineer. Apply here: \n\nVideo: \n\n![uploaded image](/media/?type=post&id=105232&key=user_uploads/3620299/89d6b5de-3729-4019-a87f-0fe27a549152)\n\n![uploaded image](/media/?type=post&id=105232&key=user_uploads/3620299/b4ac7aec-afea-4a96-9174-ae6202812c2f)\n\n", "url": "https://www.ycombinator.com/launches/RNI-rise-reforming-distributed-chemical-production", "votes": 5, "created_at": "2026-07-13T14:13:24.467Z" } ] }, { "id": 32193, "name": "Floracene", "slug": "floracene", "former_names": [ "TareRx" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/4083ec47b697a352249164c76574e273a5fa75d7.png", "website": "https://www.floracene.com/", "all_locations": "New York City, NY, USA", "long_description": "Floracene lets any healthcare operator describe the internal tool they need and deploy it in one click - connected to production data, with read and write access scoped by the user. Think Lovable or Retool, except purpose-built for the needs of healthcare business that has no engineering team.\r\n\r\nThe bottleneck in healthcare isn't building anymore. Anyone can vibecode an 80% solution in an afternoon, custom to their own workflow, with people already on staff. The tool then dies locally on the builder’s computer, because IT (correctly) won't let anyone deploy a tool touching PHI without first-class authentication and permissioning. Our first customer had multiple employees who had each built useful tools and dashboards locally, but couldn’t actually share these with coworkers or connect to their core practice data. Within weeks of implementing Floracene, they were able to consolidate their BI tooling into Floracene and shipped 4 net-new to production.\r\n\r\nA staggering ~$200B is spent on administrative software in healthcare every year in the US. Tens of thousands of those companies still run core operations on spreadsheets, and we're at the top of the first inning of AI adoption.\r\n\r\nAsaph and Jonah met on the hospitals team at Palantir two years ago, and have each spent the last four years in the plumbing of US healthcare: claims, eligibility, and payor contracting. Asaph was an early PM at Garner. Jonah was a founding engineer at Metriport (YC S22) after starting his own fax-automation startup for fertility clinics.", "one_liner": "Vibecode + deploy HIPAA-compliant internal tools, no eng team needed", "team_size": 2, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1783448694, "tags": [ "Health Tech", "Medical Devices", "Digital Health", "Enterprise Software", "Healthcare IT" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/floracene", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/floracene.json", "year_founded": 2026, "city": "New York City", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/4083ec47b697a352249164c76574e273a5fa75d7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/5ccdb13884880d40ad27da646bb9675f961c4c61.png" ], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 3605405, "full_name": "Asaph Kupferman", "title": "Founder/CEO", "founder_bio": "Founder / CEO at Floracene (YC S26). Previously Hospitals team at Palantir and early employee at Garner Health. \n\nRobotics / Electrical Engineering from Penn. Amateur chef.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aa-kup/", "x_url": "https://x.com/KupfermanAsaph", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/0a9703d8fcfaefb7b743105e311c4f6d9cc654b8.jpg" }, { "user_id": 1225540, "full_name": "Jonah Kaye", "title": "Founder/CTO", "founder_bio": "Founder / CTO at Floracene (YC S26). Previously Tech Lead at Palantir (Hospitals) and Founding Engineer at Metriport (YC S22). \n\nCS and Philosophy @ UChicago. Brazilian Jiu Jitsu Purple Belt.", "is_active": true, "linkedin_url": "https://linkedin.com/in/jonah-kaye-875a051b2", "x_url": "https://x.com/JonahKaye2", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/303600cd0aa2a45efd8dee5820b815f081eca32a.jpg" } ], "jobs": [], "news": [], "launches": [] }, { "id": 32194, "name": "GUILD", "slug": "guild", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/efd3ccc04f842f9a3a1235ba1ef66aa4447fc993.png", "website": "https://www.guildai.co", "all_locations": "New York City, NY, USA", "long_description": "GUILD is an AI-Native Defense prime making aerospace parts faster and cheaper. We win defense contracts, automate manufacturing, and deliver to the government's door.", "one_liner": "Making Defense Manufacturing Faster and Cheaper", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Defense", "launched_at": 1778869862, "tags": [ "Manufacturing", "Supply Chain", "Defense", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Defense" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": true, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/guild", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/guild.json", "year_founded": 2025, "city": "New York City", "country": "US", "linkedin_url": "https://www.linkedin.com/company/guildai/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/efd3ccc04f842f9a3a1235ba1ef66aa4447fc993.png", "app_video_url": null, "dday_video_url": "https://bookface-images.s3.us-west-2.amazonaws.com/video_event/77/1789083552-66c43d504ac5635e52a1f9823fa8ef5a/f119a8c82a28d76d89468609e5d0285b.mp4?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAQC4NIECABPQ2AHNW%2F20261020%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20261020T022739Z&X-Amz-Expires=3600&X-Amz-SignedHeaders=host&X-Amz-Signature=670ee52c4dfde4e363ba0bb12bdec7abf466e2511a9bc2abc2f9f929d0dd4501", "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/e55fc87a71d9e406a0a8fdfdd14ba685fc117785.png" ], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3161464, "full_name": "Kaya Celebi", "title": "Founder/CEO", "founder_bio": "Founder at GUILD. Previously AI Tech Lead at Hartree Partners and ML Engineer at Morgan Stanley.\nMSc in Computer Science from Columbia, dual BSc in CS and Statistics from Duke.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/kaya-celebi", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5b48b1bdd435cb7bc03969c64b95f41c522af52d.jpg" }, { "user_id": 3163115, "full_name": "Erim Gurlemis", "title": "Founder/COO", "founder_bio": "Founder at GUILD. Previously Associate in the Global Investment Office at Morgan Stanley, where he worked on wealth management development and engagement . BA in Biological Studies from Rutgers University.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/erim-gurlemis-b13b22204/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f0c16bff7dc6fd5301ab5b66684588bf18413c9f.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109739, "title": "GUILD - New Generation of Defense Contractor", "tagline": "Making aerospace and naval parts faster and cheaper for defense", "body": "Hey everyone 🙂 - we’re [@Kaya Celebi](https://bookface.ycombinator.com/user/3161464) and [@Erim Gurlemis](https://bookface.ycombinator.com/user/3163115) , co-founders of [@GUILD](https://bookface.ycombinator.com/company/32194) .\n\n**TL;DR**\n\n[@GUILD](https://bookface.ycombinator.com/company/32194) **is an AI-native defense contractor making aerospace parts faster and cheaper than today’s supply chain**. We win government contracts, turn technical requirements into production-ready workflows, and execute through factory integrations that eliminate the weeks normally lost between RFQ and factory floor. \n\n![uploaded image](/media/?type=post&id=109739&key=user_uploads/3161464/3cf82073-c0e0-4c42-8d2b-d4fab3e159ca)\n\nWe’re building GUILD in two connected motions:\n\n* First, we win and execute government manufacturing contracts directly.\n* Second, we help defense hardware companies turn designs, technical packages, and supply-chain requirements into executable procurement and production workflows.\n\nIn both cases, the goal is the same: move from technical requirements to manufactured parts faster, with less cost and less operational drag.\n\n**THE PROBLEM**\n\nGovernment contracting is fragmented, overcomplicated, and inaccessible. That makes defense manufacturing slower and more expensive than it should be.\n\nThe government waits too long and pays too much for parts the country already knows how to make. Meanwhile, the supplier base is shrinking: DoW’s reported in 2023 that small-business participation in the defense industrial base has fallen by over 40% in the last decade.\n\n**Less competition means fewer options, higher prices, weaker supply chains, and more critical work stuck before it ever reaches a production floor.** \n\n**THE CAUSE**\n\nThe defense supply chain is not wired together. Government demand, technical requirements, manufacturer capability, financial markets, and execution data all live in different places, so every contract becomes a manual search problem - stuck in PDFs, portals, spreadsheets, and domain knowledge.\n\nThe hard part is not finding government opportunities. Anyone can have Claude build a scrape-bot for [SAM.gov](http://SAM.gov) and call it a pipeline. **The hard part is turning messy technical requirements into a manufacturing decision, then executing the work.**\n\nThis favors large incumbents who can throw teams of people at the problem. Smaller manufacturers usually cannot. They either spend weeks decoding the package, price in too much risk, miss the timing, or never enter the market at all.\n\n**OUR APPROACH**\n\n**GUILD is the prime contractor**, not a marketplace or lead-gen layer for manufacturers. We win government work directly. We break the technical package into the operational steps required to produce and deliver the part. Then we execute.\n\n**The core of this is our company brain: a structural ontology of defense manufacturing.**\n\nAI-native, to us, means the company documents and learns from its own execution. Every artifact around a contract - requirements, quotes, supplier interactions, decisions, awards, workflows, and deliveries - becomes structured memory. Our agents traverse that graph to understand what a part is, how it can be made, what it should cost, and what needs to happen next.\n\nThe company brain compounds. Every bid makes the next bid smarter. Every factory handoff improves the next handoff. Every delivery teaches the system what actually happened, not just what the plan said. That is what lets us move from document parsing to execution as the contractor responsible for the work.\n\n**WHAT WE’VE BUILT**\n\nWe’ve built the company brain that turns technical packages into executable manufacturing workflows. It ingests messy contract packages, understands the requirements, prices work against known factory capability, generates the production path, and keeps improving as new parts, factories, and outcomes enter the system.\n\nThis is not a hard-coded set of workflows. The system helps us build and refine new workflows as it observes more of the defense manufacturing knowledge space. GUILD Forge is the interface for this: a procurement and production planning system for defense companies that need to move from technical design to supply-chain execution.\n\n**WHAT WE MAKE**\n\nOur current focus is aerospace precision metals. That includes jet and engine components, structural and wing-adjacent parts, housings, brackets, fittings, retainers, and the simple but critical hardware that sits inside larger aerospace systems.\n\nThese parts are not always glamorous, but they are everywhere. When their supply chain is slow, expensive, or opaque, the larger programs above them slow down too.\n\n**THE TEAM**\n\n![uploaded image](/media/?type=post&id=109739&key=user_uploads/3161464/9f3944e0-0a34-426b-a920-033542ce928f)\n\nKaya has an MS in Computer Science from Columbia and studied CS + Statistics at Duke. His background is in ML research, AI modernization, and quant work: he worked on AI modernization and quantitative systems at Morgan Stanley, and later led an AI team at hedge fund Hartree Partners. Coming into defense from the outside forced him to learn the industry from first principles and turn that learning into GUILD's technical architecture.\n\nErim studied Biological Sciences at Rutgers University, where he conducted research on metabolism before working in pharma consulting supporting oncology research. He then pivoted into finance at Morgan Stanley. There, he led full-scale AI implementation for Wealth Management & Advisory. At GUILD, he turns the messy reality of defense manufacturing, supplier coordination, pricing and execution into operating requirements our system can handle.\n\n**THE ASK**\n\nWe’d love intros to defense companies that need better access to procurement, people with government or program connections, and teams dealing with supply-chain pain around technical packages, suppliers, compliance, packaging, or delivery.\n\nWe’re especially interested in meeting defense hardware companies, aerospace manufacturers, primes, and manufacturers with end-to-end automation capability - the kind of partners where GUILD can integrate directly into the factory operating layer, not just send PDFs back and forth.\n\nEmail us at [founders@guildai.co](mailto:founder@guildai.co) or visit .\n\n\\-  Kaya + Erim 🧿", "url": "https://www.ycombinator.com/launches/SXz-guild-new-generation-of-defense-contractor", "votes": 23, "created_at": "2026-08-08T17:00:00.905Z" } ] }, { "id": 32195, "name": "RealPact", "slug": "realpact", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/0066aaeaa91920faf3f4b4ee35ce9e7c2bc99740.png", "website": "https://realpact.ai", "all_locations": "San Francisco, CA, USA", "long_description": "RealPact builds AI agents that handle the paperwork behind real estate transactions. They find property records like deeds, tax records, permits, parcel data, and MLS information, then use that data to fill out contracts, organize documents, track deadlines, and help brokerages close deals faster.\r\n\r\nWe are starting by automating transactions and will expand into other parts of brokerage operations. Much of the software used across the real estate industry is outdated. With today’s advances in AI, we can build agents that handle this operational work and create modern software around how brokerages actually work.\r\n\r\nWe currently sell to brokerages in New Hampshire and are expanding quickly.", "one_liner": "AI-native OS for Real Estate Brokerages", "team_size": 2, "industry": "Real Estate and Construction", "subindustry": "Real Estate and Construction -> Housing and Real Estate", "launched_at": 1780455733, "tags": [ "Real Estate", "Enterprise Software", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Real Estate and Construction", "Housing and Real Estate" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/realpact", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/realpact.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/realpact/", "x_url": "http://x.com/RealPact", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/0066aaeaa91920faf3f4b4ee35ce9e7c2bc99740.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 1249471, "full_name": "Ranvir Deshmukh", "title": "Founder/CEO", "founder_bio": "Previously built CourseMe, Dartmouth's AI course advisor used by 90% of students (4,000+ MAU, 1M+ page hits): Acquired.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ranvir-deshmukh-209706199/", "x_url": "https://x.com/ranvirdeshmukh_", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/87c1b073d739f5221931f4720738e22cb1ef1df1.jpg" }, { "user_id": 2392871, "full_name": "Erik Peterson", "title": "Founder/CTO", "founder_bio": "Previously built CourseMe, Dartmouth's AI course advisor used by 90% of students (4,000+ MAU, 1M+ page hits).", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/erik-peterson-mn/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/df82479ae00a881fd652cb53dc8368ea0d6e27b0.jpg" } ], "jobs": [], "news": [ { "title": "Erik Peterson ’27 and Ranvir Deshmukh ’26 win Magnuson competition with real-estate tech startup - The Dartmouth", "url": "https://www.thedartmouth.com/article/2026/05/erik-peterson-27-and-ranvir-deshmukh-26-win-magnuson-competition-with-real-estate-startup", "date": "Jun 02, 2026" } ], "launches": [ { "id": 109434, "title": "RealPact: AI-native OS for Real Estate Brokerages", "tagline": "Automating real estate transactions from the first conversation to closing.", "body": "Real estate does not have an AI problem. It has an infrastructure problem.\n\nWhile nearly every major industry is using AI to move faster, reduce manual work, and automate operations, real estate agents are still managing some of the largest financial transactions in the world through disconnected software, email threads, repetitive data entry, and paperwork.\n\nAgents should be spending their time winning clients, advising buyers and sellers, and closing deals. Instead, they lose hours gathering property information, filling out the same details across multiple forms, tracking deadlines, and coordinating every party involved in a transaction.\n\nThat is why we are building RealPact.\n\nRealPact is not another tool for completing one form or managing one part of a deal. We are building the intelligent transaction infrastructure for real estate agents and brokerages: a system that can understand a transaction, automate the paperwork, coordinate every party, track every deadline, and carry a deal from the first conversation to closing.\n\nMost real estate software stores the work. RealPact does the work.\\\n\\\n\n\nWe are already working with several major brokerages in New Hampshire and rolling out RealPact across their teams.\n\n“RealPact feels magical. We have never seen anything like it in real estate, and we believe it is going to change the industry for good.”\n\nSpenser Molloy\\\nOwner, NextHome Modern Realty\n\nOur long-term vision is simple: give every real estate agent an intelligent operating system that handles the transaction work behind the scenes, so they can spend more time building relationships, serving clients, and closing deals.", "url": "https://www.ycombinator.com/launches/ST4-realpact-ai-native-os-for-real-estate-brokerages", "votes": 8, "created_at": "2026-08-06T19:56:18.877Z" } ] }, { "id": 32198, "name": "Riften", "slug": "riften", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/d3c58cea1afce06f81a0352c8f3ecc8554d707c7.png", "website": "https://riften.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Riften builds model routing, evaluations, private models, and training environments for AI labs, AI-native companies, and enterprises.\r\n\r\nWe help teams choose models for their workloads, evaluate how they perform, and build private models around company-specific work. We’re also developing coding environments where agents learn through backend engineering tasks.", "one_liner": "Earned intelligence for every company", "team_size": 5, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1785872995, "tags": [ "Developer Tools", "Reinforcement Learning", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/riften", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/riften.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/riftenai/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/d3c58cea1afce06f81a0352c8f3ecc8554d707c7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 2941426, "full_name": "Carl Stauffer", "title": "Founder/CEO", "founder_bio": "Founder and CEO of Riften. Former Palantir Forward Deployed Engineer. Former nuclear researcher at Jefferson Lab. CS and AMS at Johns Hopkins.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/carl-stauffer-314335262/", "x_url": "https://x.com/carlstauffer_", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/0a3a2f44ecdd315423ad0b8510ef3bffafaf4843.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110330, "title": "Riften: Route every AI request. Build the models your company owns.", "tagline": "Cut LLM spend with one endpoint that selects the lowest-cost model that works.", "body": "[Riften Launch Video](https://youtu.be/9T82XHysqsc)\n\n## What Riften does\n\nRiften is an OpenAI- and Anthropic-compatible gateway that routes each LLM request to the lowest-cost model capable of completing it.\n\nConnect your existing providers, change the base URL, and keep the product you already built. Riften works across customer-facing AI products and internal tools, including Claude Code and Codex.\n\n**Riften is already removing enterprise-scale model spend from production workloads.**\n\n## Why we built it\n\nMost AI products choose a model once and send nearly everything to it.\n\nA documentation lookup, a support classification, and a codebase migration do not require the same intelligence. Yet they are routinely sent to the same frontier model at the same price.\n\nFor customer-facing products, that cost cuts directly into gross margin. For internal agents, it limits how broadly the company can deploy them.\n\n## How it works\n\nRiften evaluates each request locally, applies the company’s cost, capability, latency, and reliability constraints, and selects among the models available to that organization.\n\nHard tasks can still reach frontier models. Routine work moves to lower-cost commercial or open-weight models. Teams can keep their provider accounts, pin a specific model when needed, and inspect the cost and routing decision behind every request.\n\nThe objective is not the cheapest token. It is the lowest expected cost of completing the work.\n\n## From routing to company-owned models\n\nThe router reveals which work is repeated. Outcomes reveal whether it was done correctly.\n\nA passing test, an accepted patch, a resolved ticket, or a completed workflow can become part of a customer-specific evaluation. Riften uses those evaluations to improve routing, post-train open models around recurring work, and test those models against the APIs they are intended to replace.\n\nModels earn production traffic through the same endpoint. Frontier APIs continue to handle novel work. Repeated workloads move onto models the company controls.\n\nThat creates a direct path from inference routing to evaluation, post-training, deployment, and the applications where the work happens.\n\nThe model labs are building intelligence every company can rent. Riften is building the path to intelligence every company can own.\n\n## Team\n\nRiften is founded by Carl Stauffer, a former Palantir Forward Deployed Engineer who shipped software into the purchasing and production systems of a nine-figure manufacturer. He was Palantir’s first Neurodivergent Fellow through Alex Karp’s program, conducted nuclear research at Jefferson Lab, studied computer science and applied mathematics at Johns Hopkins, and is an a16z Scout.\n\nThe rest of the team comes from Stanford Mathematics, Berkeley computer science, and Carnegie Mellon mathematics and computer science. Their work spans Oxford OATML, SPAR, CMU CyLab, Palo Alto Networks, Millennium, Amazon Search, NASA, UCSF Health, and industrial manufacturing systems.\n\nThey have built agent orchestration platforms, LLM evaluation and fine-tuning systems, cryptographic security software, production search infrastructure, CAD-matching systems, and autonomous QA for measuring agent performance. Their research includes an ICML acceptance and current work in review at NeurIPS.\n\nThe team also brings national-level recognition in mathematics and technical research spanning manufacturing, computational science, and new energy systems.\n\nRiften combines enterprise deployment, frontier evaluation research, competitive mathematics, agent systems, security, and production ML in one team built for the full path from request to outcome to model.\n\n## The ask\n\n**We are looking for companies with meaningful LLM usage across customer-facing products or internal agents.**\n\nIf most of your traffic still runs through one frontier model, send us:\n\n* Your current default model\n* Whether the traffic is internal or customer-facing\n* Your approximate monthly model spend or volume\n\nWe will show you what should remain on the frontier, what can move today, and whether Riften is worth testing.\n\nEmail us at [info@riften.ai](mailto:info@riften.ai).", "url": "https://www.ycombinator.com/launches/ShW-riften-route-every-ai-request-build-the-models-your-company-owns", "votes": 11, "created_at": "2026-08-17T15:12:03.010Z" } ] }, { "id": 32200, "name": "Bizmark", "slug": "bizmark", "former_names": [ "LeanLog" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ee2b3b896d33b5af63938bfada70f7377592ac49.png", "website": "https://bizmark.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Hey YC! We're Bizmark. I (Rodrigo) grew up in a factory, watching my mum run everything through Excel and email.\r\n\r\nThat’s why we’ve built Otto, the OpenClaw for manufacturers.\r\n\r\nIn a few days, he learns from your systems and spreadsheets to take over repetitive work in procurement, quoting, and production planning. \r\n\r\nOtto is live at multiple factories, bringing in $108k in annual recurring revenue. Last month, he worked the equivalent of 9 full-time employees.\r\n\r\nI did CS at UChicago and built a distribution business from 0 to $60M.\r\n\r\nMy cousin and cofounder Oscar did CS at Stanford and deployed AI for the US Air Force, Amazon, and the factories of Johnson & Johnson and Intel.\r\n\r\nThere's a quarter of a million manufacturers in the US, making this a $25B market.\r\n\r\nIn 5 years, they will all run on Otto-pilot.", "one_liner": "OpenClaw for manufacturers", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1784331710, "tags": [ "Manufacturing", "Automation", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/bizmark", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/bizmark.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ee2b3b896d33b5af63938bfada70f7377592ac49.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 343702, "full_name": "Rodrigo Mosqueira", "title": "Founder/CEO", "founder_bio": "I grew one of the largest e-commerce logistics startups in Brazil from zero to $60M in revenue running 100k orders/month across 10k SKUs as the first employee. Prev. CS at UChicago, MBA from Chicago Booth.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/moag/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1ea1768af0b16fd80a12279f323138a1d48501d7.jpg" }, { "user_id": 1156158, "full_name": "Oscar Aguilar", "title": "Founder/CTO", "founder_bio": "I've implemented machine learning and optimization at Intel, Amazon, and J&J, and deployed AI for the Department of War. Studied operations Research at Stanford and Georgia Tech.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/oscar-aguilar-76bb15170/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/02e433f8d34d4f2fff110b7d3eeb44d9f36d1dba.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107941, "title": "Bizmark – Transforming physical industries", "tagline": "The operating system for companies that make, move, or sell physical goods", "body": "[Bizmark Launch](https://youtu.be/-5g3wv2OE2g)\n\nHey everyone, we're Rodrigo and Oscar, founders of Bizmark (S26).\n\n**TL;DR:** Bizmark is the operating system for companies that make, move, or sell physical goods. It connects the systems, spreadsheets, and knowledge a company already has, and turns them into optimized decisions and automated workflows: quotes priced in seconds, schedules that reflow when a machine goes down, inventory balanced between service levels and cash. Bizmark adapts to how each company operates, not the other way around. Working product in weeks, not years. Some of the most innovative American companies use Bizmark to transform their operations. Ask: intros to decision makers at manufacturers, distributors, and retailers.\n\n**The problem**\n\n> _“Quotes take too long, so I lose revenue. Or I quote too fast and give away margin.”_\n\n> _“Days of inventory are critical for my business, but I’m not sure I’m buying the right quantities at the right time, or getting the best supplier terms.”_\n\n> _“An urgent order comes in and I do not know what to run next. I’m hitting max capacity constantly, but I’m not sure I’m using my equipment and labor in the most efficient way.”_\n\nAfter 150 of these conversations, we've realized these problems share the same cause. Legacy ERPs force operators into templates that don’t reflect how the business runs, while custom implementations often run to years and seven-figure budgets. As a result, decisions that directly affect revenue, margins, and service levels are made from spreadsheets, delayed reports, and fragmented tools, with less information than the company has, often by whoever is closest, not whoever knows best.\n\n**What we built**\n\nBizmark is the operating system for companies that make, move, or sell physical goods. It connects the systems, spreadsheets, and knowledge a company already has, and turns them into optimized decisions and automated workflows. Bizmark adapts to how each company operates, not the other way around. And customers get a working product in weeks, not years.\n\n**Here’s how some of the most innovative American companies use Bizmark to transform their operations:**\n\nA spec-heavy RFQ lands in the inbox. Bizmark reads it, and a quoting engine prices it against inventory, capacity, and customer history. It returns a customer-ready quote in seconds. The team approves it, it goes out, and when the customer says yes, the order enters the system without any retyping.\n\nThen the order runs. Bizmark schedules it into your machines and crews against your objectives, using the same estimates the quote was priced on. When a machine goes down or a rush order lands, the plan reflows and the new sequence reaches the queues that run the operation. The job runs, passes quality, and ships.\n\nThe order drew down inventory. Bizmark forecasts what you will sell next, holds your days of inventory in balance – enough for minimum service levels while protecting cash – and decides what to buy, when, and from whom. It sends RFQs to qualified suppliers and compares what comes back. The team chooses, and the PO books straight into the system.\n\n**Who we are**\n\nWe’re cousins who love seeing how things get built in the real world.\n\nRodrigo grew one of the largest e-commerce logistics startups in Brazil from zero to $60M in revenue running 100k orders/month across 10k SKUs as the first employee. Prev. CS at UChicago, MBA from Chicago Booth.\n\nOscar led process optimization at Intel, Amazon, and J&J, and deployed AI for the U.S. government. Prev. Operations Research at Stanford and Georgia Tech.\n\n**Ask**\n\nIntros to decision makers at manufacturers, distributors, and retailers.", "url": "https://www.ycombinator.com/launches/S4z-bizmark-transforming-physical-industries", "votes": 8, "created_at": "2026-07-28T17:50:56.018Z" } ] }, { "id": 32201, "name": "Mosaic", "slug": "mosaic-inc", "former_names": [ "Synq Labs", "Emergent Computing" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ccea1a14d9a796c8e8670b1aa85ebb489e030d7e.png", "website": "https://mosaic.inc", "all_locations": "San Francisco, CA, USA", "long_description": "Mosaic gives teams and their coding agents shared memory across every session, so everyone can understand what happened, why decisions were made, and pick up where someone else left off.\r\n\r\nWe started as a real-time coding environment where multiple people and their agents can work together live. It went viral, defined the multiplayer AI category, and inspired frontier labs like Anthropic.\r\n\r\nTogether, we’re building the infrastructure for people and agents to share context and work together.", "one_liner": "Defining the frontier of multiplayer AI.", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1786309056, "tags": [ "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/mosaic-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/mosaic-inc.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/124373945/", "x_url": "https://x.com/mosaic_ide", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ccea1a14d9a796c8e8670b1aa85ebb489e030d7e.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 1249641, "full_name": "Shubham Patil", "title": "Founder", "founder_bio": "Co-Founder @ Mosaic", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shubhampatilsd/", "x_url": "https://x.com/shubhampatilsd", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c818f32b76cdff4a81379a63c094a3d000adc530.jpg" }, { "user_id": 2918134, "full_name": "Dorsa Rohani", "title": "Founder", "founder_bio": "Co-Founder @ Mosaic | NVIDIA, University of Toronto", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dorsarohani/", "x_url": "https://x.com/dorsa_rohani", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a0468f97f1dd1a1dfff6cfcb46a74e7a6f74a676.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111682, "title": "Mosaic — Shared memory for your team's agents", "tagline": "Sync all your teams' AI agent sessions, in one place.", "body": "**TL;DR**\n\nYour team is running dozens of agents and not one of them knows what the others did. [Mosaic](https://mosaic.inc/) syncs every AI session your team runs - Claude Code, Codex, Cursor and more - into one shared place that any agent can use. \\~2,000 installs and 37 orgs actively syncing.\\\n\\\nIt started as a real-time coding environment where multiple people and their agents work together. We [**went viral**](https://x.com/dorsa_rohani/status/2074963064231952832?s=20), defined the multiplayer AI category, and this direction has since taken off: YC added multiplayer AI to its Request for Startups, and [Anthropic is validating](https://x.com/ClaudeDevs/status/2085817074816070014) the same direction. Then we took the feature users found most valuable — centralized agent sessions — and built the product around it, continuing to lead the category ever since.\\\n\\\nIDE Demo: \\\n\\\nAgent session sync demo: \n\n### **The problem**\n\nEvery coding agent session starts from zero. No memory of what happened last session, why a decision was made, or where the last person stopped. Agents run in isolation, so the only handoff between them is ad-hoc glue that breaks the moment a team adds more agents. Teammates constantly redo work they can't see.\n\n**Mosaic** keeps the record of every agent session — what was tried, what changed, and why — and carries it from one session and one teammate to the next. Any teammate or agent picks up exactly where someone else left off instead of starting over.\n\nAs a whole, we believe the way people are thinking about multi-agent systems is wrong, and our direction is being validated in real time. Most systems copy the human org chart — specialized agents passing messages — but that structure only exists because humans can't share memory; the real unit isn't the agent, it's one shared context store every agent reads from and writes to. [Read our memo here](https://mosaic.inc/blog/memo).\n\n![uploaded image](/media/?type=post&id=111682&key=user_uploads/2918134/f7913654-964e-4579-9f66-45d531e3b009)\n\n### Who we are\n\nWe're two engineers who built and deployed multi-agent systems and hit this exact wall every time.\n\n\\- **Dorsa:** worked on frontier ML @ NVIDIA, Tenstorrent and Deep Genomics. A Transformer implementation of hers has [2,400+ stars](https://github.com/DorsaRoh/Machine-Learning) on Github. 11-time hackathon winner and one of Major League Hacking’s youngest [Top 50 Hackers](https://top.mlh.com/2023/profiles/dorsa-rohani) out of 150,000+ hackers worldwide.\n\n\\- **Shubham:** founding engineer @ Beem building multi-agent systems that reason over human intent. A serial hacker at heart, he reverse-engineered 70 campus kiosks at UCSD and chained their compute into a working distributed cloud platform.\n\n### **Our ask**\n\n* **Run multiple coding agents on your team? Try Mosaic and tell us what breaks!**\n* **Intros** to engineering teams where every developer already runs their own agent.\n* Reach us directly: [contact@mosaic.inc](https://mosaic.inc/)\n\nWe built Mosaic because we kept facing this problem ourselves. If you're living it too, we'd love to hear from you.", "url": "https://www.ycombinator.com/launches/T3K-mosaic-shared-memory-for-your-team-s-agents", "votes": 9, "created_at": "2026-08-24T14:30:00.125Z" } ] }, { "id": 32202, "name": "Dialogus", "slug": "dialogus", "former_names": [ "Zyra", "Dialogus", "Dialogus Labs" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c7df204d91879e8e730b8993b754b3d35b87106e.png", "website": "https://www.dialoguslabs.com/", "all_locations": "San Francisco, CA, USA", "long_description": "", "one_liner": "voice AI agents that learn how your company actually works", "team_size": 3, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1782528923, "tags": [ "B2B", "Customer Support" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/dialogus", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/dialogus.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/dialoguslabs", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c7df204d91879e8e730b8993b754b3d35b87106e.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 3398179, "full_name": "Juberth Rodriguez", "title": "Founder/CEO", "founder_bio": "Co-founder @ Dialogus \n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/juberth-rodriguez-398b2b277/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f6b8c1f10e8949f7df03d738ffbd42ad5324cdb5.jpg" }, { "user_id": 1603180, "full_name": "Hans Ibarra", "title": "Founder/CEO", "founder_bio": "Co-founder @ Dialogus | Ex-Google @ Gemini, Gmail and Youtube\n\nPapa Johns, KFC and Totalplay as customers", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/hans-ibarra/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/22039d81a2c60002993586e9e4fa30448494606b.jpg" }, { "user_id": 3047141, "full_name": "Rodrigo Terán", "title": "Founder/CTO", "founder_bio": "Co-founder @ Dialogus | Ex-Google, Meta | Math & Physics Olympiad", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rodrigoteranhdz", "x_url": "https://x.com/rodrigoteranhdz", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/21c36c6ec529aeff6f25beadc5c54f2946a6dab8.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107329, "title": "Dialogus: Building the infra for enterprise voice agents", "tagline": "Infra for enterprise voice agents", "body": "Every year, enterprises spend billions of dollars operating contact centers. They struggle with high labor costs, long wait times, inconsistent customer experiences, and the challenge of hiring and retaining thousands of agents.\n\nMost AI voice products try to automate individual phone calls. We believe the real opportunity is much bigger: rebuilding the contact center itself.\n\nDialogus replaces traditional contact center operations with AI voice agents that can answer calls, resolve customer requests end-to-end, integrate with enterprise systems, and scale across hundreds of locations while maintaining a consistent customer experience.\n\nToday, Dialogus is already powering AI voice operations for Fortune 500 enterprises, handling thousands of customer calls in production.\n\nThe company was founded by Rodrigo Teran, Hans Ibarra, and Juberth Rodriguez after seeing how legacy contact center software wasn't built for the AI era. Instead of layering AI on top of decades-old infrastructure, they're building an AI-native contact center from the ground up.\n\nIf you're an enterprise looking to transform your customer support with AI, we'd love to talk.\n\n", "url": "https://www.ycombinator.com/launches/Rv7-dialogus-building-the-infra-for-enterprise-voice-agents", "votes": 10, "created_at": "2026-07-24T18:40:56.654Z" } ] }, { "id": 32204, "name": "Alloovium", "slug": "alloovium", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/a18b2e1de19ebc106cb340305f6e055da094e03b.png", "website": "https://www.alloovium.com/en", "all_locations": "San Francisco, CA, USA", "long_description": "Alloovium is the document intelligence layer for construction 🏗️ It reads every document and piece of data on a project (contracts, specs, drawings, compliance records, conversations), turning them into answers you can trust and actions with real value: every answer cites its exact source sentence, paperwork is checked clause by clause against the builder's own Integrated Management System, and nothing AI-drafted moves without a human sign-off, with it all compliant by default.", "one_liner": "The company brain for construction companies", "team_size": 5, "industry": "Real Estate and Construction", "subindustry": "Real Estate and Construction -> Construction", "launched_at": 1783581007, "tags": [ "Construction", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Real Estate and Construction", "Construction" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/alloovium", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/alloovium.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "http://linkedin.com/company/alloovium", "x_url": "https://x.com/alloovium", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/alloovium", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/a18b2e1de19ebc106cb340305f6e055da094e03b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 3047254, "full_name": "Zander Schweitzer", "title": "Founder/CEO", "founder_bio": "CEO @ Alloovium (YC S26) | I'm a Civil Engineer who has spent my whole life around construction, having my first experience in a bulldozer at 5 yrs old. Since then, I've driven dozers, dump trucks and loaders, done some site surveying and cut my teeth in construction as a site engineer on some of Australia's coolest civil marine projects.\n\nNow, I'm building the software I always wanted @ Alloovium.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/zander-schweitzer-1a0651205?utm_source=share&utm_campaign=share_via&utm_content=profile&utm_medium=ios_app", "x_url": "https://x.com/zandergs15", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/957ed917a450033a2d0bfc4e96aa2004699c0638.jpg" }, { "user_id": 2271222, "full_name": "Omar Elloumi", "title": "Founder/CPO", "founder_bio": "CPO @ Alloovium (YC S26) spent three years at LexisNexis turning ML research into production systems for highly regulated legal data. \nNow i’m bringing that experience in AI & Product to build the future of construction @ Alloovium.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/omar-elloumi/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8a6445965ad14629896ff39fdd35bbafe7cf2e2f.jpg" }, { "user_id": 3650351, "full_name": "Cielo Nicolosi", "title": "Founder/CTO", "founder_bio": "CTO @ Alloovium. Coding since 8 years old, founding engineer at 4 previous startups. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/cielo-nicolosi/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/7d72e200e868bb454b82834373e3ca6254faea8a.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 105028, "title": "Alloovium - Automating compliance for major construction projects", "tagline": "Turning everything your project produces into answers you can trust and compliance you can prove.", "body": "# \\*\\*\\*\\*\n\n**TL;DR: Alloovium automates compliance for major construction projects 🚧** Construction is 13% of global GDP, and it runs on paperwork. Nobody in the industry wants another platform — they want what they already have brought together in one place. Alloovium is that unifying layer: it reads everything the project produces from the tools you already use, and turns it into cited answers you can trust and the documentation that moves the job forward — compliant with the builder's own management system, every answer cited to its source sentence, a human sign-off on anything AI-drafted.\n\n[**https://youtu.be/wVPJL3o8PzM**](https://youtu.be/wVPJL3o8PzM)\n\n**🚧 The problem**\n\nOn major projects, the record keeps losing: 80% of megaprojects blow their budget or schedule, and sums in dispute average a third of contract budgets. We think the root cause is hiding in plain sight — in the paperwork.\n\n**The root cause:** paperwork is a state in time — every document a snapshot of a project thats moving daily. The documents aren't a pile, they're a web: the safety plan leans on procedures; procedures lean on templates, registers and licences; and the tender made promises all of them have to keep true. Across every company delivering the job. People spend their days dragging the paper back to reality instead of moving the job forward.\n\n**The consequence:** one change means five documents to amend, and nobody catches the missed one until an auditor, regulator or client's lawyer does. In some jurisdictions: up to 25 years jail for directors and a $20M fine for the GC.\n\n**The cost of coping:** engineers who trained for years to build spend their days chasing paper instead. One tier-1 team is spending $2.5M and 40–50 people on a single 400-page submission of management plans — and that's before the project even starts.\n\n**The unlock:** a sleepless project sentinel, keeping watch on the documented rules, unlocking the experts to focus on the build.\n\n**🔨 What Alloovium does**\n\n* **Documents, Apps, Correspondence and Conversations centralised in the Project.** A singular place where the project becomes a reality.\n\n![uploaded image](/media/?type=post&id=105028&key=user_uploads/3047254/69afa2af-4a4e-4376-ad5e-186084b42ee4)\n\n* **Cited answers over every project document.** Click the citation, land on the exact source sentence.\n\n![uploaded image](/media/?type=post&id=105028&key=user_uploads/3047254/6687ff2d-1620-404b-b4a9-a6f8a8fd5aae)\n\n* **Routines.** Build repeatable workflows that are triggered by on-site events to always stay ahead of the game on your paperwork.\n\n![uploaded image](/media/?type=post&id=105028&key=user_uploads/3047254/ec46815c-e232-43eb-a583-ec74c80f25ff)\n\n* **Drafting from the project itself.** Call transcript to a formatted and actionable document in minutes, with AI-authorship tracking, a human still refines and reviews it.\n\n![uploaded image](/media/?type=post&id=105028&key=user_uploads/3047254/503df904-c919-46e3-9e31-c3e2c768db67)\n\n* **Continual compliance sweeps** against the builder's own IMS, clause by clause. Our new engine's first end-to-end sweep landed **39 cited findings** on one project.\n\n![uploaded image](/media/?type=post&id=105028&key=user_uploads/3047254/fd973fc4-a947-4f2a-8e9c-d2e42671b97f)\n\n* Plus meeting/inbox capture, voice to text data capture, licence-expiry alerts, and live SharePoint + Procore sync.\n\n**📈 Traction**\n\nDeploying across $400M in active projects, including a subsidiary of BMD (one of Australia's largest privately owned civil contractors). ISO compliance product in beta with a 30-year auditor as design partner.\n\n**👋 Who we are**\n\n![uploaded image](/media/?type=post&id=105028&key=user_uploads/3047254/aba0a4bb-7496-4a42-960f-bb96be1c8307)\n\n* [**Zander Schweitzer**](https://bookface.ycombinator.com/user/3047254) grew up in construction, spent his fifth birthday in the back of a bulldozer, drove dump trucks, and worked as a site engineer before quitting to start this.\n* [@Cielo Nicolosi](https://bookface.ycombinator.com/user/3650351) has coded since he was 8, was founding engineer at four startups, placed top 0.25% of his state, and won Australia's largest student hackathon.\n* Around us: an exited construction-tech founder, an ex-LexisNexis Senior Data Scientist, and a founding SDR from Askable.\n\n**⭐️ Our ask**\n\n* **Intros** to construction, engineering, or AEC, especially head contractors, GCs, or ENR Top 400 firms in the US. Construction is structurally viral (one head contractor exposes us to \\~50 subs a year), so one warm intro compounds.\n* **Feedback** if you've dealt with construction docs, compliance, or contracts.\n* **Notes** if you've sold software into archaic industries.\n\nDrop me a note at [zander@alloovium.com](mailto:zander@alloovium.com), +61 438 965 455 or direct any friends to \n\nZander & Cielo ❤️", "url": "https://www.ycombinator.com/launches/RK0-alloovium-automating-compliance-for-major-construction-projects", "votes": 8, "created_at": "2026-07-13T19:57:10.996Z" } ] }, { "id": 32205, "name": "Mochi", "slug": "mochi", "former_names": [ "Mochi", "MOCHI.TV" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/49a6b24fc91b9bae1c825be608651c20ecb90c31.png", "website": "https://www.mochi.tv/", "all_locations": "Los Angeles, CA, USA", "long_description": "We're the #1 app for AI anime, where fans make the stories. In just 2 months, we’ve grown to 20K+ creators publishing 2K+ series a month, with originals from Marvel, Toei, and Webtoon writers. To date we’ve crossed 300M views on socials.", "one_liner": "Watch and create anime", "team_size": 5, "industry": "Consumer", "subindustry": "Consumer -> Content", "launched_at": 1786602622, "tags": [ "Consumer", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Consumer", "Content" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/mochi", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/mochi.json", "year_founded": 2025, "city": "Los Angeles", "country": "US", "linkedin_url": "https://www.linkedin.com/company/mochitv/", "x_url": "https://x.com/watchmochi/", "fb_url": "https://www.facebook.com/watchmochitv", "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/49a6b24fc91b9bae1c825be608651c20ecb90c31.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/59d05b9ad2e443030795f668e6cc298ada2595dc.png", "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/a3b6abb9f5a84f70d841c70fcfca9b124bb0a7cf.png" ], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 1027100, "full_name": "Robert Qiu", "title": "Founder/CEO", "founder_bio": "uc berkeley alum, ex-faang. scaled game startup to $200k monthly revenue. launched top 50 nft project with $100m+ trade volume (justin kan, ray chan were holders).", "is_active": true, "linkedin_url": "https://linkedin.com/in/robertqiu", "x_url": "https://x.com/whatsrobdoing_", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/0c4f29fe16d966a78088fa967f48acf835bb1ec0.jpg" } ], "jobs": [ { "id": 117582, "title": "Founding Product Engineer", "url": "https://www.ycombinator.com/companies/mochi/jobs/BeGodLa-founding-product-engineer", "location": "Los Angeles, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$120K - $180K", "equity_range": "0.25% - 0.75%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [ "React Native", "TypeScript", "Figma", "AI Agents" ] } ], "news": [], "launches": [ { "id": 110404, "title": "MOCHI.TV - 1 Minute Anime", "tagline": "A mobile app to watch vertical anime shorts, created with AI.", "body": "**TL;DR:** [MOCHI.TV](http://MOCHI.TV) is a mobile app to watch vertical anime shorts. Users can scroll to discover shows and binge bite-sized episodes, unlocking them with either coins or a subscription. We produce new shows daily with AI, and they've hit 300M+ views organically.\n\n\n\n**The problem**\n\nTraditionally, entertainment takes years to go from production to the TV screen. But fans always crave more. With the advent of AI, entertainment cycles become as fast as days, and now we’re able to serve the most underserved audience of all - anime fans.\n\n**What we built**\n\nA mobile app to watch 1 minute anime packed with:\n\n* emotional thrills (_finally_, you get revenge on the bully), \n* plot twists (_turns out_, the bully is your secret long-lost sibling), \n* and cliffhangers (_suddenly_, the bully whispers “without me, they’ll come for you…”)\n\nUnlike traditional anime, our content is built for Gen Z. TikTok doomscrolling and lower attention spans has led to audiences wanting payoffs in seconds instead of seasons. Our AI workflow allows us to drop a new series a day and build a daily habit with our users, with shows written by top writers from Toei Animation, Spider-man, Webtoon, Spongebob, Robot Chicken, and more.\n\nAnd users are not just passive viewers either - they can create their own through the app as well, with remixing and creator monetization on the way.\n\n**Traction**\n\n300M+ views, 100K+ users, and 900K+ followers across socials, grown organically.\n\n\n\n**Looking ahead**\n\nWith AI, we’re able to build durability like a media company, but iterate like a software company. Making hits is no longer a creative risk but a matter of time and data. Anime is our beachhead, but we plan to expand to every vertical - from live action to dating shows, to k-dramas and kids TV - with AI at the forefront. Like Pixar in the 90s, or Netflix in the 00s, we see an opportunity to build the next great entertainment company on the cusp of a new technology.\n\n**Team**\n\n* **Robert Qiu:** Scaled game startup to $200K/mo, UC Berkeley alum and ex-FAANG\n* **Michael Ninh:** Watched 400+ anime, led consumer growth at Hush (S16) and Massdrop, managed 100s of video creatives\n\nWe're lifelong anime fans with a love for consumer products - this is the most exciting thing we’ll ever build.\n\n**Our ask:** Watch anime (or create one) at \n\nReach us at [founders@mochi.tv](mailto:founders@mochi.tv)", "url": "https://www.ycombinator.com/launches/Sii-mochi-tv-1-minute-anime", "votes": 46, "created_at": "2026-08-14T20:36:04.113Z" } ] }, { "id": 32206, "name": "DeepReach Inc.", "slug": "deepreach-inc", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/165caaebecbd3f0b9470bb5817a0a1817520c935.png", "website": "http://www.deepreach.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Physical AI — robots and world models — is bottlenecked on one thing: diverse,\r\nreal-world data at massive scale showing how humans actually do work in the\r\nphysical world. It was never written down, and you can't buy it, because no one\r\nhas ever captured it. Only real local experts, the people who do this work every\r\nday, can produce data this authentic and valuable.\r\n\r\nDeepReach builds the network that reaches them. We design and manufacture\r\nwearable stereo capture devices, then place them with local data partners who\r\nrun collection as their own business — in warehouses, workshops, farms, kitchens\r\nand repair shops. They own the local relationships; we supply the hardware, the\r\nsoftware platform, the quality pipeline and the payments. Anyone can run a data\r\nbusiness.\r\n\r\nThat structure is what produces diversity. Instead of thousands of hours inside\r\none environment, we deliver hours spread across thousands of environments —\r\nwhich is what actually makes robot policies generalize.\r\n\r\n475 devices in the field and 100+ data partners are live across 7 countries, adding new\r\npartners every week. They've collected nearly 150,000 clips in the past three\r\nmonths, with volume roughly doubling month over month. Our data is in production\r\nwith multiple frontier model and robotics companies, and we're scaling to\r\n10,000+ devices across 50 countries.", "one_liner": "Network for Physical AI Data Business", "team_size": 6, "industry": "B2B", "subindustry": "B2B", "launched_at": 1786563914, "tags": [ "Hardware", "Machine Learning", "Robotics", "Computer Vision", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": [ { "question": "How did your company get started? (i.e., How did the founders meet? How did you come up with the idea? How did you decide to be a founder?)", "answer": "Chris and I are from the same hometown and have known each other for ten years, long before either of us thought about robots.\r\n\r\nIn that time I built two companies around the same problem: how do you organize people you'll never meet, in places you'll never visit, to do consistent work? The first was HireIO, a cross-border staffing platform out of Silicon Valley. The second was Talex, an expert network for LLM data annotation — around 70,000 people across ten countries. Chris went the other direction: a CS PhD at USC in computer vision, ML infrastructure at Meta, then founding ML engineer at Aven.\r\n\r\nRunning an annotation network for language models, I kept hitting the same wall from the supply side. Text was running out, and the next generation of models wasn't going to learn from the internet. It was going to learn from the physical world — and nobody had captured it. What a warehouse picker, a line cook or a phone repair technician knows is real expertise, and none of it was ever written down. You can't buy that data, because it doesn't exist yet. Chris was hitting the same wall from the model side, watching what was missing from training sets.\r\n\r\nThe default answer at the time was to put a lab in one city and record thousands of hours in the same room. That produces volume, not diversity, and robot policies don't generalize from volume. What the models needed was the opposite shape: a few hours each from thousands of different real workplaces, all over the world. Which meant the bottleneck wasn't an AI problem. It was a distributed human network problem — reaching thousands of ordinary businesses across dozens of countries, and giving the people already working in them a reason to capture what they do. That was the problem I'd spent two companies learning to solve.\r\n\r\nWe started DeepReach in 2025. Chris owns perception and the data engine; I own the network, hardware operations and the commercial side. Ten years of knowing each other means we skipped most of the negotiation that takes co-founders their first year." }, { "question": "What is the core problem you are solving? Why is this a big problem? What made you decide to work on it?", "answer": "Core problem: robot models can't generalize, because the data they learn from all looks the same.\r\n\r\nPhysical AI needs to see how work actually happens — a mechanic torquing a bolt, a baker shaping dough, a technician opening a phone. None of it was ever written down. The internet captured what people say about the world; it never captured how people act in it. So unlike language models, this generation can't be trained on scraped text. Every hour has to be collected on purpose. \r\n\r\nToday the field has roughly tens of thousands of hours of this data against a need estimated in the hundreds of millions. But the gap isn't really a volume gap. It's a diversity gap. The standard approach is to build a lab, hire operators, and record for months in one controlled space. That yields thousands of hours that all look alike, and a policy trained on it fails the moment the lighting, the counter height, or the tool changes. Our customers' acceptance specs now reflect this directly: they cap how much of a delivery can come from any single environment type, and reject batches that are too concentrated. Diversity is the binding constraint, and it's now contractual.\r\n\r\nDiversity is hard for an economic reason, not a technical one. Collecting one hour each from a thousand real workplaces means negotiating access to a thousand real businesses — each with an owner who has no reason to let a stranger film their operation, in a country where you have no presence. That cost is why everyone else builds the lab instead. It's a distribution problem wearing an AI problem's clothes. \\\r\n\\\r\nThat's why we built the company as a network rather than an operation. The people who already work in those places, and already have the trust and the access, run collection as their own local business. We give them the hardware, the software and the payments. Access stops being something we buy and becomes something the network already has. \\\r\n\\\r\nI decided to work on this from the supply side. I'd spent two companies building cross-border human networks — a staffing platform, then a 70,000-person annotation network across ten countries — and kept watching demand arrive for data that no existing structure could produce. The bottleneck was never the model. It was that nobody had built the network that reaches the physical world." }, { "question": "What is your long-term vision? If you truly succeed, what will be different about the world?", "answer": "The long-term vision is a network that can reach any workplace on earth, and a record of how human work is actually done.\r\n\r\nConcretely: 10,000+ devices, 50 countries, one million commercial scenes, ten million task types. Every deployment adds to what we call the Human Skill Library — a permanent record of how people do physical work, built from the inside by the people who do it. Much of this knowledge exists nowhere else. It lives in the hands of a mechanic who has been doing the job for thirty years, and when he retires it's gone. Text and video on the internet captured what people say about work. Nobody has captured the work. \r\n\r\n\\\r\nIf we succeed, two things are different：\\\r\nFirst, robots stop being confined to the places that can afford to build around them. Today automation only pays at Amazon scale — a facility large enough to redesign itself for the machine. That's a rounding error of the world's actual workplaces. A policy trained on a million real environments doesn't need the environment to change; it works in a repair shop with bad lighting and a counter at the wrong height. That's the difference between robotics as a capability the largest companies buy and robotics as infrastructure a small business can use.\\\r\nSecond, the people whose work teaches these models are inside the economy that results. The last generation of AI was trained on data scraped from people who were never asked and never paid. This one has to be collected on purpose, from real workers in real workplaces, which means it can be built the other way — as something people opt into and earn from. We already have data partners in ten countries running this as a business they own. At 10,000 operators that isn't a side effect of how we source data. It's the point." } ], "url": "https://www.ycombinator.com/companies/deepreach-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/deepreach-inc.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/in/timothyli-deepreach/", "x_url": "https://x.com/DeepReach_AI?s=20", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/deepreach-ai", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/165caaebecbd3f0b9470bb5817a0a1817520c935.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/482534d82410b1adf1774b961acd50c6704093e0.png" ], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3114257, "full_name": "Chris Liu", "title": "Founder/CTO", "founder_bio": "Co-founder at DeepReach Inc. Founding MLE & tech lead of Aven scaling to #1 in DM volume in the US. Research Scientist at Meta working on core ML systems. PhD in computer vision from USC. B.S. from Peking University.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/drxcliu", "x_url": "https://x.com/drxcliu", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1f9e9212493e67c1bb9a3a913e2ab4c6fa9b897e.jpg" } ], "jobs": [ { "id": 112083, "title": "VP / Head of Global Data Partner Network & Operations", "url": "https://www.ycombinator.com/companies/deepreach-inc/jobs/r0itCEs-vp-head-of-global-data-partner-network-operations", "location": "San Jose, CA, US", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "$150K - $200K", "equity_range": "0.50% - 1.00%", "experience": "3+ years", "visa": "Will sponsor", "skills": [] } ], "news": [], "launches": [ { "id": 110389, "title": "DeepReach: Launch a real-world data business today", "tagline": "Multi-million-dollar contracts secured with frontier AI labs — and growing", "body": "**TL;DR:** DeepReach is a global data network that equips local entrepreneurs with wearable devices to capture the real-world operation data for Physical AI training. By decentralizing collections across thousands of entrepreneurs who have access to local businesses, they provide high-diversity training data that cannot be accessed elsewhere. DeepReach already signed multiple-million-dollar contracts with frontier AI labs.\\\n\\\n**DeepReach is building a global network for real-world human data to train Physical AI.**\n\nWe enable local entrepreneurs to operate data-collection businesses using our wearable capture devices, while we provide the hardware, software, QA, and customers.\n\nThe result is a distributed network that produces the diverse human demonstrations Physical AI labs cannot obtain anywhere else. Every new data partner expands the diversity and value of the network.\n\nIn just three months, DeepReach has helped **150+ entrepreneurs** build data businesses employing **1,000+ local experts** across **7 countries**. Together, they’ve collected **500,000+ real-world operation clips**, captured **150,000+ professional skill demonstrations**, and secured **several million dollars in customer contracts** with frontier AI labs and robotics companies.\n\n▶ **Watch how it works:** \n\n---\n\n# **The Problem**\n\nFrontier AI is already learning from the physical world.\n\nThe challenge is no longer collecting data, it’s collecting **diverse** real-world data at scale.\n\nMost companies solve this by building collection labs or hiring workers directly. Both approaches scale headcount, but neither scales access. The hardest part is reaching millions of real workplaces across different industries, countries, and working conditions.\n\nWithout that diversity, AI struggles to generalize beyond a small number of controlled environments.\n\n---\n\n# **What We Built**\n\nInstead of building another data collection company, **we built a platform where entrepreneurs launch independent real-world data businesses.**\n\nDeepReach provides the wearable devices, software platform, quality infrastructure, customer demand, operational support, and payments. **Entrepreneurs build and grow the businesses.**\n\nThey recruit local experts, develop trusted relationships with local businesses, and continuously expand into new industries, communities, and real-world environments.\n\n**Every new business expands the platform’s reach into the physical world, creating access to places, people, and skills that centralized data collection simply can’t reach.**\n\nThe result isn’t a larger workforce, it’s **a growing ecosystem of independent data businesses** that continuously expands the world’s supply of diverse real-world data for Physical AI.\n\n---\n\n# **Where We Are**\n\nIn just three months:\n\n* **150+ entrepreneurs** have launched data businesses on DeepReach\n* Those businesses now employ **1,000+ local experts** across **7 countries**\n* **500,000+ real-world operation clips** collected\n* **150,000+ professional skill demonstrations** captured\n* **Several million dollars in signed customer contracts** with frontier AI labs and robotics companies\n* Scaling toward **10,000+ devices** across **20 countries**\n\n# **Where This Goes**\n\nOur model is simple: labs list the tasks and environments they need, our entrepreneurs pick up the orders and go record them, and every delivery makes the network more capable of fulfilling the next one. It also compounds into something bigger, what we call the **Human Skill Library**: a growing record of how people interact with the physical world.\n\nMost of this knowledge has never been documented. It lives in the hands of experienced workers, passed down through practice rather than words, and it often disappears when they retire.\n\nOur vision is to preserve and scale humanity's physical knowledge, so every AI system can learn from the collective experience of human work.\n\n# **About Us**\n\n\\\nTim built HireIO into a leading global workforce company with 15 millions in annual revenue, then founded [Talex.ai](http://Talex.ai), an AI-powered expert network, before starting DeepReach. Chris comes from a computer vision background. We discovered the problem firsthand while labeling data for robotics startups: the real bottleneck wasn’t annotation, it was capturing diverse real-world human skills at scale.\n\n![uploaded image](/media/?type=post&id=110389&key=user_uploads/1294315/9bb8c293-ab57-42aa-b756-58562e546b78)\n\n---\n\n# **How You Can Help**\n\nIf you’re building frontier AI systems and need more diverse real-world data, we’d love to talk.\n\nIf you’re an entrepreneur interested in building a real-world data business, applications are open here: [www.deepreach.ai/join](https://www.deepreach.ai/join)\n\nTim Li\\\nFounder & CEO, DeepReach AI\n\n[**tim.li@deepreach.ai**](mailto:tim.li@deepreach.ai)", "url": "https://www.ycombinator.com/launches/SiT-deepreach-launch-a-real-world-data-business-today", "votes": 470, "created_at": "2026-08-13T18:33:15.556Z" } ] }, { "id": 32220, "name": "COACH", "slug": "ai-coach", "former_names": [ "COΛCH" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/78b58d361b887533d84ef36e56f833ff7fd87c5c.png", "website": "https://www.getcoach.com/", "all_locations": "San Francisco, CA, USA", "long_description": "85% of all sales still happen face-to-face. Coach is how those get closed.\r\nOur AI turns in-depth sales conversation analysis into company knowledge and personalized coaching to help reps close more deals.", "one_liner": "The AI Coach for in person sales reps.", "team_size": 4, "industry": "B2B", "subindustry": "B2B -> Sales", "launched_at": 1785353697, "tags": [ "SaaS", "B2B", "Sales", "Sales Enablement", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Sales" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/ai-coach", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/ai-coach.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/my-aicoach/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/78b58d361b887533d84ef36e56f833ff7fd87c5c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 1619431, "full_name": "Mathieu Perez", "title": "Founder/CEO", "founder_bio": "Founder & CEO at COΛCH. \n\nStudied CS at EPFL, UPenn, and U Sydney.\nPreviously co-founded Alsace Test, scaling to €3.5M revenue in 14 months at age 18. \nFormer professional goalkeeper for Atlético Madrid and Racing Club de Strasbourg.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/mathieu-perez-719019201/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fd9e1e0d6345383f6a93dc854c0bba5f488fa9c6.jpg" }, { "user_id": 1977183, "full_name": "Thomas Perez", "title": "Founder/COO", "founder_bio": "Co-founder & COO of COΛCH (YC S26)\n\nI run customer success and our forward-deployed approach: 100% pilot conversion and 0% churn.\n\nBefore:\nStudied Business at Audencia & USYD. Coding at Le Wagon. \nCo-founded Alsace Test, scaling to €3.5M revenue in 14 months at age 21.\nEx-football player — Atlético Madrid & Torcy academies, UA (NCAA).\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/thomas-perez-13011/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/557b87d34999f30ffbc51a8bd2ef13b3803af1d1.jpg" }, { "user_id": 2929268, "full_name": "Nicolas Sebag", "title": "Founder/CTO", "founder_bio": "Founder & CTO at COΛCH\nWe transform real-life conversations into actionnable intelligence.\nSpent the last 10 years building software products in hospitality, aeronautics, healthcare and more", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/nicolas-sebag-504747118/", "x_url": "https://x.com/sebcny", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d51cd4f662c1032baa37d98bc2a71fa8b58497ed.jpg" } ], "jobs": [ { "id": 111820, "title": "Founding AI Engineer", "url": "https://www.ycombinator.com/companies/ai-coach/jobs/VHrbq7B-founding-ai-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "AI engineer", "salary_range": "$150K - $250K", "equity_range": "0.50% - 1.50%", "experience": "6+ years", "visa": "US citizen/visa only", "skills": [ "PostgreSQL", "Python", "Machine Learning", "AI Agents" ] }, { "id": 111851, "title": "Founding Designer", "url": "https://www.ycombinator.com/companies/ai-coach/jobs/zQZFaPA-founding-designer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Design", "role_type": "Product design", "salary_range": "$120K - $180K", "equity_range": "0.25% - 1.00%", "experience": "6+ years", "visa": "US citizen/visa only", "skills": [ "Design systems", "Figma" ] }, { "id": 112049, "title": "Founding Customer Success (Senior)", "url": "https://www.ycombinator.com/companies/ai-coach/jobs/Y46rQ4Y-founding-customer-success-senior", "location": "Paris, France", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "€55K - €70K EUR", "equity_range": "0.05% - 0.25%", "experience": "3+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 112050, "title": "Founding Customer Success", "url": "https://www.ycombinator.com/companies/ai-coach/jobs/Y3oskDm-founding-customer-success", "location": "Paris, France", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "€45K - €65K EUR", "equity_range": "0.01% - 0.15%", "experience": "Any (new grads ok)", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 115448, "title": "Founding Sales - Europe", "url": "https://www.ycombinator.com/companies/ai-coach/jobs/HE1Omqf-founding-sales-europe", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "€60K - €120K EUR", "equity_range": "0.05% - 0.30%", "experience": "3+ years", "visa": "US citizenship/visa not required", "skills": [] } ], "news": [], "launches": [ { "id": 108339, "title": "COACH: The AI coach for in-person sales reps", "tagline": "Make every rep sell like your best one: door-to-door, in-home, showroom.", "body": "Hey everyone, we're Mathieu, Thomas and Nicolas, two brothers and a best friend building COACH.\n\n**TL;DR:** We're building the best coach for in-person sales ever built: reps record every in-person meeting from our app, and our AI turns it into precise, personalized coaching. And it works so well that 5 months in, thousands of reps already run on COACH, and we're live with the top 5 in-person sales companies in France, lifting conversion rates by 32% on average.\n\n\n\n## The Problem\n\n85% of sales in the world happen face-to-face. And in-person selling is brutally hard: a rep alone in someone's living room, reading signals, handling doubts, feeling the exact moment to push.\n\nThe best reps catch what no playbook can teach. That's why 20% of them bring in 80% of the revenue, and nobody knows what they're doing differently. The only way to find out is to physically be there: a ride-along, a full day, 100 km of driving to watch a single meeting. It doesn't scale.\n\n## The Solution\n\nCOACH's AI learns how to close customers for your business, and closes the gap at every layer of the sales org:\n\n* **Every rep gets a personal coach:** they record every in-person meeting from our mobile app, in-home, showroom, door-to-door, and walk out of each one with personal, precise feedback and a clear path to the next close. Our AI learns from your top 20% and coaches everyone to perform like them, starting each day with a morning brief on their weakest point. 90% improve on it in their first week.\n* **The company gets a sales brain that learns how to close its customers:** every meeting feeds it. Why is one region closing twice as much as another? Which objection is killing deals this month? Leadership just asks, COACH has the answer.\n\nThousands of reps already run on COACH, live with the top 5 in-person sales companies in France, where we've lifted conversion rates by 32% on average, and dozens of reps 3–4×'d their numbers. We're now growing fast in the USA, where our first customers are switching from virtual ride-along platforms to COACH.\n\nEvery rep gets a coach. Your company gets a brain.\n\n## The Team\n\nAt 18 and 21, we built Alsace Test: a 44-rep field sales team that did €3.5M in 14 months. Alsace made us fall in love with in-person sales, obsessed, actually.\n\nAnd one thing was driving us crazy: same product, same material, same training, and one guy closed 4× more than the others. We were driving 2,000 km a week just to watch reps sell.\n\nSo we asked them to record themselves. Suddenly, without moving, we were listening to 10 hours of meetings a day. That was COACH before COACH.\n\nIn an AI-driven sales world, we're deeply convinced Zoom sales and call centers will soon be replaced by AI agents. Face-to-face selling will be the last human job on Earth. We're betting on it.\n\n## Our Ask\n\nIf you run 5+ field sales reps, let's talk, we'll show you how to make them superhuman.\n\n* **Book a demo:** [Calendly](https://calendly.com/mathieu-perez-ai-coach/nouvelle-reunion)\n* **Email me:** [mathieu@getcoach.com](mailto:mathieu@getcoach.com)\n* **Text/call me:** +1 628-358-7470", "url": "https://www.ycombinator.com/launches/SBP-coach-the-ai-coach-for-in-person-sales-reps", "votes": 51, "created_at": "2026-07-30T19:46:52.311Z" } ] }, { "id": 32221, "name": "Grocalo", "slug": "grocalo", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/0fdeca47d8a703741544113635c43a9f56c783f7.png", "website": "https://grocalo.com/", "all_locations": "Los Angeles, CA, USA", "long_description": "Grocalo (S26) is an AI brain that runs content for creators. We understand their audiences, using video understanding and deep platform analytics to bottle up the taste behind each creator and personalize and distribute content to creator specific audiences.\r\n\r\nGrocalo generated 2B views and gained 4M followers for our creators in the last 2 months. Some of our clients include Jimmy O. Yang (Comedian, 11M followers), King Bach (72M followers), Leenda Dong (28M followers), Steven He (30M followers), Sheena Melwani (18M followers), and Jeannie Mai (TV host, 6M followers).\r\n\r\nContent should be personalized and data-driven. We’ve developed the most advanced AI brain replicating viral content for any given audience. \r\n", "one_liner": "AI brain for creators that runs their social content", "team_size": 2, "industry": "Consumer", "subindustry": "Consumer -> Content", "launched_at": 1787022465, "tags": [ "Entertainment", "Creator Economy", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Consumer", "Content" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/grocalo", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/grocalo.json", "year_founded": 2026, "city": "Los Angeles", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/0fdeca47d8a703741544113635c43a9f56c783f7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 3672474, "full_name": "David Hwang", "title": "Founder/CEO", "founder_bio": "Founder at Grocalo. Spent 5+ years in Investment Banking covering Enterprise Software and Internet & Digital Media and the last 4 years working with my partner to grow her content end-to-end, including generating 10B views and gaining 5M followers. BA in Economics from Johns Hopkins.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/david-hwang-b5301436/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/eb67ea357271a98b7f1d4c207e4d8f0206185662.jpg" }, { "user_id": 208755, "full_name": "Alex Yee", "title": "Founder/CTO", "founder_bio": "Founder, Grocalo. 2x YC founder, built & sold onederful (W18). Founding engineer at Lyra Health ($5B+). MS BioE, Stanford; BS, UCSD.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/alexyee1/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/94f0c17bfc19684d4c5c368f26fe48f9fcf46be2.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111327, "title": "Grocalo: AI brain that runs content for creators", "tagline": "Running socials for top content creators and celebrities", "body": "![uploaded image](/media/?type=post&id=111327&key=user_uploads/208755/f101d76e-515f-4146-8bcd-78f1d81b4726)\n\n**What is Grocalo?** \n\nGrocalo (S26) is an AI brain that runs content for creators. We understand their audiences, using video understanding and deep platform analytics to bottle up the taste behind each creator and personalize and distribute content to creator specific audiences.\n\nWe’re developing the most advanced AI brain capable of replicating virality for any given audience.\n\n**We run content for top entertainers**\n\nGrocalo generated 2B views and gained 4M followers for our creators in the last 2 months. Our clients include Jimmy O. Yang (Comedian, 11M followers), King Bach (72M followers), Leenda Dong (28M followers), Steven He (30M followers), Sheena Melwani (18M followers), and Jeannie Mai (TV host, 6M followers).\n\n**Our team**\n\nWe have been best friends since middle-school. Alex is a 2nd time YC founder who built and sold Onederful (YC W18). David spent 5+ years in investment banking and for the last 4 years has been working with his partner to grow her content end-to-end, including generating 10B views and gaining 5M followers.\n\nEmail [team@grocalo.com](mailto:team@grocalo.com) to learn more.", "url": "https://www.ycombinator.com/launches/Sxb-grocalo-ai-brain-that-runs-content-for-creators", "votes": 9, "created_at": "2026-08-20T19:30:00.407Z" } ] }, { "id": 32222, "name": "Audun", "slug": "audun", "former_names": [ "Solvva" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/e7da249d05c3b78e6493d6546cc70a78c450280f.png", "website": "https://audun.co", "all_locations": "Oslo, Oslo, Norway", "long_description": "Audun automates debt collection.", "one_liner": "AI-native debt collection company", "team_size": 5, "industry": "Fintech", "subindustry": "Fintech", "launched_at": 1786093910, "tags": [ "Fintech", "Finance", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech" ], "regions": [ "Norway", "Europe" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/audun", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/audun.json", "year_founded": 2026, "city": "Oslo", "country": "NO", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/e7da249d05c3b78e6493d6546cc70a78c450280f.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 3587634, "full_name": "Jacob Hambro", "title": "Founder/CEO", "founder_bio": "Founder at Audun. Previously consultant at McKinsey & Company. Holds a MSc in Computer Science & Finance from NTNU.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jacob-hambro-70a9b1153/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5fd98a82faf0f9b01a9675f4b4cbce1e3332c7b7.jpg" }, { "user_id": 3632034, "full_name": "Magnus Vonen", "title": "Founder", "founder_bio": "Founder at Audun (YC S26), building AI-native debt collection. Previously at McKinsey & Company, advising on go-to-market strategy and AI. Before that, four years as an operator in Norwegian special forces, including overseas deployments. Economics at NHH (Norwegian School of Economics)", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/magnus-nj%C3%B8lstad-vonen-518945208", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/dcea88c387054dd5fdfdfacef04c7a032a94d818.jpg" }, { "user_id": 3649218, "full_name": "Henrik Brun Fevang", "title": "Founder", "founder_bio": "MSc in Computer Science and Artificial Intelligence. Previously a full-stack software engineer and architect at PwC, building software infrastructure and data platform for SINTEF, one of Europe’s largest research institutions. Previous full-stack engineering experience at Oda and Airthings.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/henrik-brun-fevang-8b5b18195", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/46d31d5e3c94ede241e440fdc92dff2df6c88006.jpg" }, { "user_id": 3414987, "full_name": "Peter Flo", "title": "Founder & Chief Scientist", "founder_bio": "Founder & Chief Scientist at Audun. Previously Special Projects at Midjourney and a graduate researcher at MIT. MS in Data Science from Harvard, dual B.S. from UC Berkeley. More at pflo.org.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/peterhollevik", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6cde847c6e23c09b48938d4f381e4b99258eeffe.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111293, "title": "Audun: AI-native debt collection", "tagline": "Full-stack AI debt collection company", "body": "A letter arrives in the mail.\n\nImpossible to understand.\n\nAnother one.\n\nThen the phone calls start.\n\nOnce a day. Unpleasant, sometimes hostile.\n\nThis is how debt collection works in 2026.\n\n**We’re here to fix it.**\\\n\\\n[https://drive.google.com/file/d/1X9mVp2ZyJMUhCcHN6NadzMNLV5S63Wpv](https://drive.google.com/file/d/1X9mVp2ZyJMUhCcHN6NadzMNLV5S63Wpv/view?usp=sharing)\n\n# TL;DR\n\nAudun is an AI-native debt collection company. Our agents automate the collections process end to end, from reminders and negotiations to back-office and legal work. This reduces collection costs for creditors, and enables humans to spend their time on the conversations that matter.\n\n![uploaded image](/media/?type=post&id=111293&key=user_uploads/3587634/839f47e4-cdd7-45ed-befd-c621930fef65)\n\n# Ask\n\nWe’re looking for introductions to collection leaders at banks, lenders, utilities, telcos, and other high-volume creditors.\n\n# The problem\n\nThe debt collection industry is broken. It runs on call centers, paper letters, and software built in the 90s. While the rest of the world is preparing for AGI, the topic of the industry is still email and cloud.\n\nAll these inefficiencies ultimately end up with the consumer: incomprehensible letters arriving by snail mail, hostile phone calls, and fees stacking up on a bill they may never have seen. They're paying the cost of an archaic system.\n\n# The solution\n\nAudun is an AI-native debt collection company. We automate up to 90% of cases across the whole business, from front to back office. Our agents manage reminders, negotiations, payment plans, and settlements across voice, SMS, email, and WhatsApp. They operate within creditor-defined rules, log every action, and escalate sensitive cases to humans with the full context attached. Behind the scenes, they handle tasks such as skip tracing, payment verification, and legal document drafting.\n\nBy automating the grunt work of collections, Audun lowers collection costs for creditors and, ultimately, borrowing costs for consumers. We enable people to focus on the conversations that matter, where human judgment is still required, giving customers the treatment they deserve.\n\nWe started out in the Nordics, which has some of the strictest collections regulation in the world. Compliance is critical in this industry, so we built our agents to handle the toughest version of it first. \n\n# Who we are\n\nWe’re four founders with experience across McKinsey, Midjourney, and Norway’s first unicorn. We’ve seen firsthand how manual this industry is, and how stuck the incumbents are. We decided to build the company this industry needs from the ground up.\n\n![uploaded image](/media/?type=post&id=111293&key=user_uploads/3587634/fbfdfe9e-490e-4603-ad97-15b8ead694a6)\n\n", "url": "https://www.ycombinator.com/launches/Sx3-audun-ai-native-debt-collection", "votes": 8, "created_at": "2026-08-24T16:00:00.405Z" } ] }, { "id": 32223, "name": "Marble", "slug": "joinmarble", "former_names": [ "Remy", "Truffle" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/fd8594643506090563d23eac2d8bee3b9c1e526f.png", "website": "https://joinmarble.ai/", "all_locations": "New York City, NY, USA", "long_description": "Marble is the first AI-native operating system for the restaurant back-of-house. \r\n\r\nRestaurants, cafes and bars still run on a patchwork of disconnected point solutions, spreadsheets, clipboards and tribal knowledge, which costs the industry billions each year in stockouts, food waste, and lost profit. \r\n\r\nMarble replaces that entire stack with one system. Computer vision and voice AI capture real-time inventory levels 90% faster than manual counts, ML models forecast item-level demand at 97% accuracy, and AI agents run dozens of end-to-end workflows autonomously, from procurement and invoice reconciliation to scheduling, food prep, menu engineering, and more.", "one_liner": "The autonomous back-of-house for restaurants.", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1782079217, "tags": [ "SaaS", "Restaurant Tech", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/joinmarble", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/joinmarble.json", "year_founded": 2026, "city": "New York City", "country": "US", "linkedin_url": "https://www.linkedin.com/company/joinmarble/", "x_url": "https://x.com/joinmarble", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/fd8594643506090563d23eac2d8bee3b9c1e526f.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/b4010ede5729d3391d58ce8364837483ffe142b9.png" ], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 1278164, "full_name": "Aakar Khanna", "title": "Founder/CEO", "founder_bio": "Co-Founder and CEO at Marble (YC S26). Previously Head of Strategy & Operations at Kouper (healthcare applied AI) and an early BizOps hire at AtoB (YC S20). Started career as an Investment Banking Analyst at Goldman Sachs covering Restaurants and Food & Beverage. Grew up managing family's multi-unit restaurant business. Undergrad from the University of Michigan's Ross School of Business with a minor in Computer Science. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aakarkhanna/", "x_url": "https://x.com/khannauno", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/10b55e1df529bf42691cc6106a60970b8dcc3cc2.jpg" }, { "user_id": 306060, "full_name": "Arjun Chaliha", "title": "Founder/CTO", "founder_bio": "Co-Founder & CTO at Marble (YC S26). Previously Senior Software Engineer at Bloomberg, leading AI/ML initiatives for Electronic Trading. CFA Charterholder. BSE Computer Science from University of Michigan. While in college, founded a peer-to-peer tutoring marketplace, backed by University of Michigan.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/achaliha/", "x_url": "https://x.com/arjunjchaliha", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/08d4fd14f0ed0a02d3cd986bfeaf28c11c0ee14a.jpg" } ], "jobs": [ { "id": 109611, "title": "Founding Engineer", "url": "https://www.ycombinator.com/companies/joinmarble/jobs/JIr3duM-founding-engineer", "location": "New York, NY, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$150K - $225K", "equity_range": "1.00% - 3.00%", "experience": "1+ years", "visa": "US citizen/visa only", "skills": [] }, { "id": 110135, "title": "Founding Ops", "url": "https://www.ycombinator.com/companies/joinmarble/jobs/1V1GfLm-founding-ops", "location": "New York, NY, US", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "$100K - $150K", "equity_range": "0.20% - 0.75%", "experience": "1+ years", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 108273, "title": "Marble: The autonomous back-of-house for restaurants", "tagline": "Marble uses computer vision and voice AI to count inventory in minutes and a collection of AI agents that run inventory and labor management on autopilot.", "body": "Hey YC! We’re Arjun and Aakar, and we’re launching Marble (previously Truffle), the first AI-native operating system for the restaurant back-of-house.\n\n**TL;DR:** We help restaurants control their two biggest costs, inventory and labor—saving them thousands of dollars per month in food waste, stockouts, and labor overage. We replace today's patchwork of disconnected point solutions, spreadsheets and clipboards with a single, easy to use system that actually does the work. We use computer vision and voice AI to count inventory 90% faster than manual counts, ML models to forecast item-level demand with 97% accuracy, and AI agents that run dozens of end-to-end workflows autonomously, including ordering, invoice reconciliation, scheduling, food prep, menu engineering, and more.\n\n**The result**: We're live and scaling rapidly across multi-unit operations in the US and Canada. You can find us today in A&W Restaurants, The Kati Roll Company, Sophie’s Cuban, Kokomo NYC, HolyShred, Dead Letter No. 9 and many more. We’ve driven a 40% reduction in food waste, eliminated stockouts, and saved managers 30+ labor hours per week.\n\nCheck us out here:[ www.joinmarble.ai](http://www.joinmarble.ai)\n\n**Our ask:** If you run or know someone who runs a restaurant, café, bar, or hospitality group, we'd love to talk. You can reach us at [founders@joinmarble.com](mailto:founders@joinmarble.com).\n\n**Everyone knows someone who runs a restaurant, bar, or cafe - we’re also giving away $500 (or an Oura Ring) per referral!** [www.joinmarble.com/refer](http://www.joinmarble.com/refer) \n\n \n\n**The Problem**\n\nRestaurant operators face an uphill battle. Ultra-competitive markets, rising food and labor costs, unforgiving hours, and after all of it, they take home 5% in net profit (if they're lucky).\n\nA restaurant can be full all day and night, but it's the back office operations that determine whether that money hits the bottom line. Today that back office still runs on clipboards, spreadsheets, and guesswork. Inventory gets counted by hand, if at all, so no one has a real-time picture of what's actually on the shelf. Orders get placed from memory, so operators either over-buy (contributing to $150B+ in annual food waste) or under-buy and stock out on the exact items guests came for. Food prep and schedules are built on best-effort guesses, and invoices go unchecked while suppliers quietly raise prices line by line.\n\nEvery one of these is a margin decision, and every one is being made blind, leading to errors that compound every day, across every location.\n\n**The Solution**\n\nMarble runs the entire back office for restaurants. This centers around managing the two biggest cost centers (inventory and labor) and works in 4 phases:\n\n* **Capture:** Computer vision + voice AI through our mobile app allow operators to count inventory 90% faster than manual counts, with 99% accuracy. No more hours spent counting items in a walk-in freezer with a clipboard (IYKYK). Invoices and deliveries are captured the same way.\n* **Forecast:** ML models predict item-level demand for the weeks ahead at 97% accuracy. In addition to that, LLM-powered agents scan the signals a spreadsheet never sees - local events, weather, news, reservations, holidays - and adjust the forecast for what's actually happening around each location.\n* **Act:** A set of AI agents run the workflows that actually move the needle on profitability: placing optimized orders, building demand-informed prep sheets, drafting schedules, proactively tracking and flagging waste, reconciling invoices, engineering menus, and more.\n* **Learn:** This industry runs on tribal knowledge. We have built the product as a closed-loop system from day one: every count, correction, and order flows back into the platform as context that our agents carry forward, so Marble learns how each location actually runs, adapts to its patterns automatically, and gets sharper every day.\n\n![uploaded image](/media/?type=post&id=108273&key=user_uploads/1278164/1dd8a396-a754-4277-ad3b-1613bf1d4212)\n\n**Team**\n\nWe’ve been close friends for nearly 10 years, since our freshman year at Michigan, and we're having a blast taking on this industry.\n\nAakar grew up managing the back office for his family's multi-unit quick service restaurants, where he saw the problem up close. He later covered the restaurant sector as an investment banker at Goldman Sachs before leaving to lead growth at an applied-AI startup in healthcare.\n\nArjun is a second-time founder and was previously a Senior Software Engineer at Bloomberg, where he led AI/ML initiatives on the Electronic Trading team. At Bloomberg, he built out complex trading systems that processed tens of millions of dollars daily.\n\n![uploaded image](/media/?type=post&id=108273&key=user_uploads/1278164/2bab9f3a-cca8-4846-8be1-6cf2a4e38dcf)\n\n", "url": "https://www.ycombinator.com/launches/SAL-marble-the-autonomous-back-of-house-for-restaurants", "votes": 40, "created_at": "2026-07-30T17:31:56.143Z" } ] }, { "id": 32225, "name": "Stoa", "slug": "stoa", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/cf71f2bfca9f3f554d22c43f0dfd7b1a38171535.png", "website": "https://www.stoaexchange.com", "all_locations": "San Francisco, CA, USA", "long_description": "Stoa provides pricing and trading infrastructure for AI hardware, live with an RFQ marketplace where companies buy and sell GPUs through vetted dealers real-time.", "one_liner": "The market behind AI hardware.", "team_size": 3, "industry": "Fintech", "subindustry": "Fintech", "launched_at": 1785367811, "tags": [ "Fintech", "Hardware", "Marketplace", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/stoa", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/stoa.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/stoamarkets/", "x_url": "https://x.com/stoaexchange", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/stoa-d396", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/cf71f2bfca9f3f554d22c43f0dfd7b1a38171535.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 1961995, "full_name": "Berat Celik", "title": "Founder/CEO", "founder_bio": "CEO of Stoa. Repeat founder who walked away from xAI to build Stoa. UBC, computer science and economics; Cornell M.Eng., computer science.", "is_active": true, "linkedin_url": "https://linkedin.com/in/beratcelik1", "x_url": "https://x.com/beratcelik0", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3db0925f484f9622bdabb784771b1618d0408f66.jpg" }, { "user_id": 1960190, "full_name": "Eren Berke Saglam", "title": "Founder/CFO", "founder_bio": "CFO of Stoa. Traded interest rates derivatives at Citi. Dartmouth salutatorian, computer science and economics.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/eren-berke-saglam", "x_url": "https://x.com/erenberke0", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8fa08c8d9b7dd4e7053173162485cba755b2b252.jpg" }, { "user_id": 3415052, "full_name": "Kaan Yigit", "title": "Founder/CTO", "founder_bio": "CTO of Stoa. Worked as a quantitative developer at Uniper, building the trading systems behind financial oil and gas books. UIUC, computer science.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/kaan-yigit/", "x_url": "https://x.com/kaan_yigit0", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/94061540990806801bdde87e10a5a8fad1228468.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108473, "title": "Stoa: The market behind AI hardware", "tagline": "GPU marketplace where vetted dealers bid blind with firm quotes. $300M+ in RFQs in our first month.", "body": "**TL;DR:** AI hardware trades over phone calls and forwarded spreadsheets. No venue, no price to mark against: lenders mark billions in GPU collateral off appraisals. We're building Stoa, an RFQ marketplace where buyers and sellers get firm quotes within 48 hours, and settled trades build the prices the market can mark, finance, and hedge against.\n\n\n\nGPU deals close in group chats. We know because we've been in them.\n\nA price arrives as a voice note. Someone forwards an inventory list, then it gets forwarded again, and soon three brokers are offering the same machines. Nobody knows who actually owns what or whether any quote is firm.\n\nThe market behind those chats is worth hundreds of billions of dollars a year. Lenders mark billions in GPU collateral off appraisals, not trades.\n\nWe're Berat, Eren, and Kaan. We're building Stoa.\n\nAI hardware is becoming an asset class, but its market is missing the basics: a venue where hardware can trade, prices that buyers, sellers, and lenders can mark against, and eventually financing and hedging built on those prices. We're starting with the venue: an RFQ marketplace for GPUs. It gets buyers and sellers firm quotes, and every bid and settlement builds the history the rest needs.\n\n**Here's how it works**\n\nPost what you want to buy or sell: B300s, H200s, H100s, a bare card or a full cluster, new or secondary.\n\nVetted dealers bid blind with firm quotes. Instead of spending weeks calling around, you get through the RFQ in 48 hours.\n\nWe handle the unglamorous parts too: KYB and counterparty checks, contracts, shipping, and settlement.\n\nIn our first month, $300M+ in RFQs ran through the book.\n\n**Why the marketplace comes first**\n\nAsk what an H100 is worth and everyone you call gives you a different answer.\n\nOn Stoa, dealers quote against the same RFQ without seeing each other's bids. Because every RFQ, firm quote, award, and settlement runs through us, we see what the hardware actually trades for.\n\nA quote is an opinion. A settled trade is a receipt.\n\nThat transaction history becomes prices buyers, sellers, and lenders can mark against. Once marks are grounded in settled trades, financing and hedging can follow.\n\nPlatts started printing oil prices in a magazine. Kelley Blue Book began as one dealer's published list of what he'd pay for cars. GPU inventory is still being passed around in forwarded spreadsheets, and Blackwell upgrades are now pushing H100 fleets into the secondary market.\n\nSomeone is going to be the price.\n\n**The team**\n\nWe've known each other for more than ten years. We have founded companies, traded interest rate derivatives, and built trading and pricing infrastructure for oil and gas. Now we're building it for AI hardware.\n\n**Buying or selling?**\n\nIf you're buying GPUs or compute, or you're sitting on hardware, sign up at [stoaexchange.com](http://stoaexchange.com).\n\n**Schedule a call: [https://stoa.cal.com/stoa/intro](https://stoa.cal.com/stoa/intro)**\n\n**Email us: [founders@stoaexchange.com](mailto:founders@stoaexchange.com)**\n\n**Text or call us:** +1 510 605 3844\n\nAnd if you know infrastructure heads at neoclouds or data centers, hyperscaler capex teams, dealers, ITADs, or credit desks lending against GPUs, please send this their way.\n\nBerat, Eren & Kaan", "url": "https://www.ycombinator.com/launches/SDZ-stoa-the-market-behind-ai-hardware", "votes": 779, "created_at": "2026-07-31T12:13:42.502Z" } ] }, { "id": 32226, "name": "Glen", "slug": "glen", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/cc11fde647ded2f52bca6ff53c7ad045f290e1ee.png", "website": "https://www.tryglen.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Glen is the layer that unifies every agent your company runs. Glen pulls in agent sessions from Claude Code, Codex and Cursor, plus Slack messages, GitHub prs, tickets, docs and calls. When a new agent picks up a task, Glen finds relevant prior context and injects it in the agents context window to steer it, producing knowledge work that is significantly higher quality, faster and cheaper.\r\n\r\nGlen does more. You can query the original agent behind any line of code, or search every session your org has ever run. It distills what your team does into skills and surfaces the right one automatically. Agents work multiplayer, picking up the style and context of whoever else is on the file. Transcripts are unified across Claude Code, Codex and Cursor, so you can hand a session from one to another mid-task. And Glen can also double as a company brain, from Slack, in app, or by just installing Glen on any agent.\r\n\r\nWe’re building toward a future where memory is no longer the missing piece of intelligence. A future where every agent maintains a lossless and continuous thread of institutional experience enabling the era of the compounding organization and fully unsupervised digital labour.", "one_liner": "Institutional Learning Layer for Every Agent in Your Company", "team_size": 1, "industry": "B2B", "subindustry": "B2B -> Productivity", "launched_at": 1781114514, "tags": [ "Generative AI", "B2B", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Productivity" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/glen", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/glen.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/tryglen", "x_url": "https://x.com/try_glen", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/cc11fde647ded2f52bca6ff53c7ad045f290e1ee.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3237090, "full_name": "Nikos Dritsakos", "title": "Founder/CEO", "founder_bio": "Founder at Glen. Ex Head of Special Projects (MTS) at Composio. Before that VP of Technology at FliteHouse at 22, post-acquisition of SalesBop (his first company, which he bootstrapped to acquisition for $500K in 18 months). At FliteHouse, scaled the eng team, led SOC 2 compliance, and helped grow to one of the biggest players in the voice agents for Canadian auto dealers space via exclusive partnerships across Canadian dealer associations. CS from Brock University.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/nikos-dritsakos/", "x_url": "https://x.com/nikos1", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/bb5fa61b15c1eadc35aa8dfce11f70397440ddde.jpg" } ], "jobs": [ { "id": 108585, "title": "Founding Member of Technical Staff", "url": "https://www.ycombinator.com/companies/glen/jobs/QiR0kCz-founding-member-of-technical-staff", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$170K - $250K", "equity_range": "0.50% - 3.00%", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [ "React", "TypeScript" ] } ], "news": [], "launches": [ { "id": 111187, "title": "Glen: A shared learning layer for all the agents in your organization", "tagline": "The shared learning layer for agents.", "body": "Hey everyone! I'm Nikos, the founder of Glen.\n\n**TL;DR**\n\nGlen is building the organizational learning and context layer for agents.\n\nEvery agent session your team runs produces reasoning worth keeping, and nearly all of it is discarded the moment the window closes.\\\n\\\n[https://drive.google.com/file/d/1T0CkjYfo3k3FB40_R7UPNfONO-SdgfKB](https://drive.google.com/file/d/1T0CkjYfo3k3FB40_R7UPNfONO-SdgfKB/view?usp=sharing)\n\nGlen is the layer that unifies every agent and model your company run. Glen pulls in agent sessions from Claude Code, Codex and Cursor, plus Slack messages, GitHub prs, tickets, docs and calls. When a new agent picks up a task, Glen finds relevant prior context and injects it in the agents context window to steer it, producing knowledge work that is significantly higher quality, faster and cheaper than ever before.\n\nAnd Glen does more. It unifies all of your org's context: query the original agent behind any line of code, search every session your org has ever run, dynamic skill distillation, multiplayer agents, unified transcripts you can hand between Claude Code, Codex, and Cursor mid-task, and a company brain in Slack.\\\n\\\nComposio scaled usage from 2 users to org wide. Over a dozen SF startups use Glen. I’m onboarding new teams every week, email me at [nikos@tryglen.com](mailto:nikos@tryglen.com) to get set up with early access.\n\n![uploaded image](/media/?type=post&id=111187&key=user_uploads/3237090/2d8c514f-6318-491a-937e-87f1d3e32ae6)\n\n**What Glen does practically**\n\n* Relevant context injected on every prompt which steers your model to do knowledge work more efficiently saving your time, money and headaches. These injections can consist or niche knowledge unique to the repo you are working in, high level understanding of relevant sub systems, and previous approaches teammates took to similar problems. Glen surfaces the prior work that matters automatically, so tasks finish faster and cost fewer tokens with no loss in quality, because what lands in the window is the relevant work rather than more of it. You can lazy-prompt instead of retyping the same background every session. So far, internal evals show about a **29% reduction in token spend** and **21% reduction in task completion time**.\n\n \\\n **Also…**\n* PR review with the original context. Highlight any code in a diff and ask why it's there. Glen queries the agent session that produced it and answers from the actual trace instead of guessing from the diff.\n* Code search across every agent trace. Ask your coding agent anything about the codebase and it searches every session any agent in your org has run, so the answer comes from the work that produced the code.\n* A skill library the whole org shares. Glen distills what your team does into skills, publishes them where everyone can use them, and surfaces the right one automatically when someone hits a task it covers.\n* Multiplayer agents. You and your designer work the same deck in parallel, you on copy and them on design. Make a design edit and your agent applies the style theirs has learned. When they touch the copy, their agent writes in the voice yours was taught. Neither of you hands anything off.\n* No lock-in to a model provider. Transcripts from Claude Code, Codex, and Cursor all land in one place, and you can hand a session from one to another mid-task, with Glen converting the format so the next agent picks up exactly where the last one stopped. Your context, your skills, and your shared artifacts come with you, so moving between providers costs nothing.\n* We have a Slackbot company brain too!\n\n**Deployed org-wide at Composio**. Composio was my first design partner. It started with two users, grew to about twenty naturally as I iterated on the product, and is now used across the company. Also in use at over a dozen other companies!\n\nIn the last 30 days alone, Glen injected **10s of millions of** characters of organizational context into their agents. These injections steer the models used by these teams, giving them a competative edge that allows them to perform knowledge work better, faster, and cheaper.\n\n**My Ask**\n\nIf you are at a startup and are reading this, there's a 80% chance Glen would be a great addition to your productivity stack. Most teams run agents heavily, and if yours is tired of re-explaining meetings, Slack, tickets, conventions, and what past sessions already worked out, that's exactly what Glen removes. \\\n\\\nI'd love to chat, and I'll come to your office, or wherever you are, and set you up for a with a early access workspace before public availability. Just shoot me an email at [nikos@tryglen.com](mailto:nikos@tryglen.com)!", "url": "https://www.ycombinator.com/launches/SvL-glen-a-shared-learning-layer-for-all-the-agents-in-your-organization", "votes": 11, "created_at": "2026-08-20T17:30:00.860Z" } ] }, { "id": 32227, "name": "Context.dev", "slug": "contextdev", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/fe20b06c889ed9f1a1b7d9b0ecf522c8e057d68b.png", "website": "https://context.dev", "all_locations": "San Francisco, CA, USA", "long_description": "Context.dev gives AI agents structured data from any website through a single API.\r\n\r\nWe help developers build smarter agents & products that depend on accurate, fresh web data.\r\n\r\nIf you need structured data from the internet, you need Context.dev\r\n\r\nProud to power agentic products at Mintlify, Super, Vizzy, Klarna, and 600+ other companies. ", "one_liner": "We give AI agents realtime web context via API", "team_size": 6, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1780323126, "tags": [ "B2B", "AI", "APIs" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/contextdev", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/contextdev.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/contextdev/", "x_url": "https://x.com/getcontextdev", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/fe20b06c889ed9f1a1b7d9b0ecf522c8e057d68b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 2538368, "full_name": "Yahia Bakour", "title": "Founder/CEO", "founder_bio": "Founder at Context.dev, building the future of web context for AI.\n\nPreviously Principal SWE/Manager at Sunrun. Ex-Amazon SDE2, tech lead on tier-1 retail checkout platform. Co-founder at StockAlarm.io (acquired), scaled to 225K+ users, 100K+ daily notifications.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/yahia-bakour/", "x_url": "https://x.com/mynameisyahia", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6b419df2397373e6789300f64b85611798b1b6f2.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109242, "title": "Context.dev: Live web data API for AI agents", "tagline": "One API to collect, structure, and continuously refresh internet data for AI", "body": "Hello!\n\nYahia here, founder of [Context.dev](https://link.context.dev/yc) (YC S26).\n\n**TL;DR:** [Context.dev](https://link.context.dev/yc) turns the live web into structured data for AI agents and applications. Trusted by Mintlify, Passionfroot, Rho, and more than 400 customers.\n\nThe web was built for humans to browse, not agents to understand.\n\nOnce an AI product needs live internet data, its team often ends up maintaining browsers, proxies, crawlers, parsers, queues, retries, caches, and multiple data vendors.\n\nWe built Context.dev so that entire workflow starts with one API.\n\n\n\n**How it works**\n\nGive Context.dev the sources your product cares about and tell us what context you need. We collect it, structure it, and keep it current without your team building the infrastructure underneath.\n\n* Send URLs, domains, companies, websites, sitemaps, or files.\n* Get back clean Markdown, screenshots, company data, or structured JSON matching your schema.\n* Use Batches to process large workloads and Monitors to keep information current.\n* We handle the browsers, proxies, rendering, extraction, caching, retries, and queues underneath.\n\nHere’s what customers are already using it for:\n\n* **Soria** monitors hundreds of healthcare and financial sources.\n* **Murph** structures data on 2M+ foods, supplements, and tests.\n* **Mintlify** turns GitHub repositories into custom branded documentation.\n* **Executor** discovers and structures API docs across 3,000+ integrations.\n* **Dench** monitors social platforms for leads, generating $3M+ in pipeline.\n* More stories at [context.dev/customers](http://context.dev/customers)\n\nYour team builds the product. Context.dev keeps it connected to the live web.\n\n**The team**\n\nI’m Yahia Bakour, founder of[ Context.dev](https://link.context.dev/yc). Before this, I led engineering at Sunrun, worked at Amazon, and built and sold two startups.\n\nI started [Context.dev ](https://link.context.dev/yc)after watching teams repeatedly rebuild, and get stuck maintaining, the same web data infrastructure.\n\nToday, our team of 4 is building the context layer between the internet and every agent or application that needs to understand it.\n\nModels are no longer the main bottleneck for intelligent knowledge work. Context is.\n\n**Our ask to you**\n\nIf you’re building an AI product that needs internet data at scale, **please reach out**. We would love to learn what you’re building and where the existing tools fall short.\n\nYou can also try the API with 500 free credits and no credit card at [**Context.dev**](https://link.context.dev/yc)\n\nSchedule a call: \\\nEmail: [yahia@context.dev](mailto:yahia@context.dev)", "url": "https://www.ycombinator.com/launches/SPy-context-dev-live-web-data-api-for-ai-agents", "votes": 113, "created_at": "2026-08-05T18:02:48.007Z" } ] }, { "id": 32230, "name": "Codag", "slug": "codag", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/0a9882f0ac224b2b32ff0636a2c28986acf45c0d.png", "website": "https://codag.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Codag compresses tool call outputs before agents read them, which reduces token spend by 20-30% while retaining output accuracy.\r\n\r\nGovern and see analytics on agent dollars across all model providers and internal agents. \r\n ", "one_liner": "The AI gateway for cheaper, faster, permissioned agents.", "team_size": 1, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1779593681, "tags": [ "Developer Tools" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/codag", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/codag.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/0a9882f0ac224b2b32ff0636a2c28986acf45c0d.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 2706494, "full_name": "Michael Zhou", "title": "Founder/CEO", "founder_bio": "Solo @Codag. ex. @Okta, @Shopify for infrastructure. BSc in Computer Science and Neuroscience from University of Toronto. Ex. Rank 80 Challenger LoL, Rank 200 TFT", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/mzxzhou/", "x_url": "https://x.com/michaelzixizhou", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/083c434760353beee5fa8582901b2fa08a1757c5.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110923, "title": "Codag: Compression and control for agent tools", "tagline": "Optimize agent costs without changing how developers work.", "body": "Hi everyone,\n\nI’m Michael, founder of Codag.\n\n**TL;DR:** Codag compresses tool results before models read them, lowering costs and improving accuracy without changing how developers work, while providing tool usage analytics.\n\n\n\n### The problem\n\nAgents consume huge amounts of output from searches, tests, builds, logs, file trees, and APIs.\n\nA single tool result can add thousands of mostly repetitive lines to a session. The model processes that output before provider-native compaction can help. Important evidence gets buried, so the agent searches again, retries, or rereads the same data.\n\nThis makes agent sessions slower and more expensive. Companies also lack a consistent view of which tools and MCP servers agents use across Claude Code, Codex, and other harnesses.\n\n### What Codag does\n\nCodag attaches to existing agent harnesses with one setup command.\n\nIt handles each result according to the action being performed:\n\n\\- Tests preserve failures, traces, file locations, and exit status.\n\n\\- Searches preserve exact file and line references.\n\n\\- Logs preserve chronology, causal evidence, and outliers.\n\n\\- Anything omitted remains selectively retrievable.\n\nOptimized tool results are 75% smaller on average, while preserving signal.\n\nDevelopers keep working normally. Codag reduces what reaches the model before the model pays to read it.\n\nCodag also reports which tools and MCP servers agents invoke, along with tokens, model-adjusted spend, savings, reducer cost, and latency across providers. Because Codag already sits in the tool path, it becomes the natural place for companies to control which tools agents can use, under what policies, and at what cost.\n\n### Results so far\n\nMore than 90 organizations use Codag, including teams of up to 35 people. I built and sold the product solo.\n\n### Where this goes\n\nOur bet is that agent activity will grow faster than model prices fall. As companies give agents more tools, longer tasks, and production access, they will need infrastructure that reduces, measures, and controls agent tool use across every provider.\n\nCodag starts with compression. The larger opportunity is becoming the gateway through which companies understand and control how agents use tools.\n\n### The ask\n\nIf your team uses agents, check out [codag.ai](http://codag.ai), or contact me at [michael@codag.ai](mailto:michael@codag.ai)", "url": "https://www.ycombinator.com/launches/Sr5-codag-compression-and-control-for-agent-tools", "votes": 6, "created_at": "2026-08-16T18:44:00.099Z" } ] }, { "id": 32233, "name": "Exosat", "slug": "exosat", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c71c0447005f2a40e7e7610af283842dbce65e39.png", "website": "https://www.exosat.com", "all_locations": "San Francisco, CA, USA", "long_description": "Exosat is building thousands of low-cost communication satellites to provide a neutral, sovereign alternative to Starlink. By designing our satellites with zero ITAR/EAR-controlled content, we can leverage manufacturers anywhere in the world to produce our satellites at under a tenth of the cost and launch them from other nations to bypass capacity bottlenecks and access equatorial orbits. \r\n\r\nWe're starting off by selling connectivity to device OEMs and network operators, but will eventually use the distributed compute across our constellation to run low-latency inference for robots and autonomous vehicles in remote and rural areas anywhere on the planet.\r\n\r\nWe've already signed hundreds of millions in LOIs to work with operators across Central Asia, Southeast Asia, and the Middle East, booked our first satellite launch next year, and hold an 11,000 satellite filing with the ITU to build the largest comms network outside of the U.S. and China.", "one_liner": "Building a neutral, sovereign alternative to Starlink", "team_size": 3, "industry": "Industrials", "subindustry": "Industrials -> Aviation and Space", "launched_at": 1786056981, "tags": [ "Space Exploration", "Satellites", "Aerospace", "Telecommunications", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Aviation and Space" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/exosat", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/exosat.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/exosatorbit/", "x_url": "https://x.com/Exosat_", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c71c0447005f2a40e7e7610af283842dbce65e39.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 689300, "full_name": "Edward Ge", "title": "Founder/CEO", "founder_bio": "Founder @ Exosat. 2x space-tech founder. Previously co-founded Aethero building radiation-hardened GPUs for edge computing and satellite autonomy. Launched 2 satellites (including the first Nvidia Orin in orbit and another one in partnership with Booz Allen Hamilton). Grew it to millions in revenue from big primes like Blue Origin and AFRL.\n\nBuilt CubeSats and high-powered rockets since high school. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/edwardge9/", "x_url": "https://x.com/EdwardGe", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2ceae5edec26bbe78bb8df0aabb142b12e6e4cf9.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109889, "title": "Exosat — Building a sovereign Starlink", "tagline": "Building thousands of low-cost communication satellites as a neutral, sovereign alternative to Starlink. We launch the first one next April.", "body": "**TL;DR: Exosat is building a sovereign alternative to Starlink with thousands of low-cost satellites. We’ve booked our first launch next April and signed hundreds of millions in LOIs with customers across six nations. We’re looking to connect with operators interested in sovereign networks and AI infrastructure. Email us: [founders@exosat.com](mailto:founders@exosat.com)**\n\nExosat is building thousands of low-cost communication satellites to provide a neutral, sovereign alternative to Starlink. \n\nWe launch our first satellite **next April**, have signed **$400M in LOIs** with customers across 6 nations, and hold an ITU filing for 11,000 satellites to build the largest comms network outside of the United States and China.\n\n[Video](https://youtu.be/aABRZfcYAFQ)\n\n### **⏰ Problem**\n\nNations and enterprises are wary about depending on Starlink.\n\nThese concerns aren’t without merit. In 2022, Starlink access was suddenly cut to stop a Ukrainian counteroffensive. As a result, nations have realized that a foreign superpower owning your satellites makes you ultimately dependent on their continued permission to access it.\n\nIn Europe, telecoms are warning about the danger of relying on foreign-owned AI and satellite infrastructure:\n\n![uploaded image](/media/?type=post&id=109889&key=user_uploads/689300/9dc3272a-bf85-4710-ad4d-e4592100ca1e)\n\nIn Indonesia, telecoms are lobbying for a Starlink ban:\n\n![uploaded image](/media/?type=post&id=109889&key=user_uploads/689300/5596ba12-4d01-4acd-b11c-8b1e777bb663)\n\nIn India, Starlink rollout has been delayed due to security concerns:\n\n![uploaded image](/media/?type=post&id=109889&key=user_uploads/689300/e4e7cd9a-e67d-42f8-bd8a-c1a78a0f4034)\n\nAnd even nations like Canada are moving off of dependence on American space infrastructure:\n\n![uploaded image](/media/?type=post&id=109889&key=user_uploads/689300/b73fbed9-3742-4105-8a86-de42843dbd1b)\n\nBut even if they don’t trust Starlink, that doesn’t mean they necessarily trust constellations like Qianfan or Guowang that are owned and operated by Chinese state enterprises.\n\nHowever, other alternatives outside of the U.S. and China face years of delay or need millions of dollars just to build a single satellite, which leaves hundreds of millions people still without reliable connectivity. \n\n### **🛰️ Our Solution**\n\nExosat fixes this problem by building thousands of satellites at a tenth of the usual cost.\n\nWe focus entirely on civil and commercial customers outside of the United States or China and fly our satellites in lower orbits versus incumbents like AST and Iridium, enabling us to take advantage of better link budgets, faster latency, and a milder radiation environment.\n\nThese factors let us use global manufacturing hubs or local partners to build our satellites using automotive-grade parts and **zero** ITAR/EAR-controlled supply chains, enabling us to seamlessly launch from other nations to bypass capacity bottlenecks and access equatorial orbits.\n\nAs a result, we can deploy a communications satellite with a 3-5 year lifespan for **\\~$300k instead of $5M** and in days instead of months, enabling us to manufacture and deploy a mega-constellation at near cost-parity to Starlink.\n\n**Our thesis is simple: we’re building a satellite at <10% of the cost to dominate 80% of the market instead of going after everyone.**\n\n### **👋 About Us**\n\nI’m a second-time space founder who previously built Aethero, a company developing satellite edge computers. We scaled to millions of dollars in revenue from defense primes like Blue Origin and Booz Allen Hamilton, won multiple SBIR awards, launched two satellites to orbit, and did radiation research on GPUs with the Aerospace Corporation and AFRL.\n\nI started Exosat and moved to Singapore because I think a future where SpaceX and China monopolize the space economy is a bleak one. I’m willing to go for broke, and put down the deposit for our first launch slot next year with my own money.\n\n### **📈 What have we done?**\n\nSince starting out this summer, we’ve:\n\n* **Booked our first launch** for next April, with early pilots underway to start initial commercial service for satellite IoT no later than the end of 2028.\n* **Added $400M+ in LOIs** from early customers across Southeast Asia, the Middle East, and Central Asia to either lease future network capacity from our dedicated constellation or build and launch their own sovereign networks with our satellites.\n* **Published ITU filings to build an 11,000 satellite constellation** to build the biggest satellite network outside of the United States and China.\n\n![uploaded image](/media/?type=post&id=109889&key=user_uploads/689300/2420ddf1-1da9-4f74-b6ef-15dd51f24eb2)\n\n### **🙏 Our Ask**\n\nIf any of these describe you or your organization, we'd love to talk:\n\n* Network operators interested in space-based capacity\n* National space agencies interested in sovereign space capabilities\n* Industrial users in maritime, energy, mining, logistics, who rely on satellites for tasks such as AIS or remote connectivity\n* OEMs producing satellite-connected devices and gateways\n\n👉 Reach us at [info@exosat.com](mailto:info@exosat.com)\n\n**Fun fact:** We’re the first YC space company to be based in Asia.", "url": "https://www.ycombinator.com/launches/SaP-exosat-building-a-sovereign-starlink", "votes": 29, "created_at": "2026-08-10T20:30:00.246Z" } ] }, { "id": 32253, "name": "Touchmark", "slug": "touchmark", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/fac4fb7e32c7424e6eea88690415f386be2f5832.png", "website": "https://touchmark.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Buy tokens over a time window from the provider of your choice at a price fixed today - 10-30% off the on-demand price. If your plans change - resell your reserved capacity to another buyer.", "one_liner": "Market for future inference capacity", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1781312801, "tags": [ "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/touchmark", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/touchmark.json", "year_founded": null, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/touchmark-ai", "x_url": "https://x.com/TouchmarkAI", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/fac4fb7e32c7424e6eea88690415f386be2f5832.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 1998938, "full_name": "Ilia Bolgov", "title": "Founder/CEO", "founder_bio": "Co-founder of Touchmark. Product @ Revolut, Maths @ Imperial College London.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ilia-bolgov-8683b0254", "x_url": "https://x.com/iliabolgov", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3922ff830a5e34f8d4d4a2a31ed11113a28e954e.jpg" }, { "user_id": 3532865, "full_name": "Roman Yanushevskyi", "title": "Founder/CTO", "founder_bio": "Co-Founder at Touchmark. Previously 2x Quantitative Research intern at Citadel Securities, AI Engineering intern at Lovable. Math and CS graduate from the National University of Singapore. IOI gold medalist.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/roman-yanushevskyi/", "x_url": "https://x.com/ryanushevskyi", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c1c4b14a1f4d9113b9badc24e14f63df1cde9840.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110492, "title": "Touchmark: Market for future inference capacity", "tagline": "Buy tokens in advance with flexibility to resell unused capacity up to 30%+ off - or more for larger orders via direct quotes", "body": "**TL;DR:** We are building a market for future inference capacity. You buy tokens for a model over a time window from the provider of your choice at a price fixed today (up to 30% off - or more for larger orders via direct quotes). If your plans change, you will be able to resell your reserved capacity to another buyer.\n\n\n\n**Problem**\n\nToken spend is growing and it’s growing fast. Out of 100+ companies we spoke to, most are trying to cut their inference costs.\n\nReserved capacity guarantee deals provide discounts + certainty, but are negotiated bilaterally, take time, and are **hard to switch from** - as a buyer you commit by prepayment, which you lose if you switch to a different provider, or in some cases even a different model.\n\nWhat if a buyer could:\n\n1. See all possible discounted capacity deals in one place\n2. Buy with a click of a button\n3. Have an option to resell to another buyer with another click of a button\n\n**But why would model providers (sellers) want this?**\n\nIn short:\n\n1. Model providers’ goal is to maximise GPU utilization - the more information they have about future demand, the better they can allocate capacity.\n2. Selling future capacity is an excellent way to finance GPU compute - dilution-free and interest-free\n3. Extra distribution channel for enterprise buyers\n\n**What if a new model comes out but I already bought capacity for the older one?**\n\nContracts for “Best model in class” allow you to buy frontier model capacity in advance. And even if you're holding a contract on an older model when a new generation ships - far from every application needs the frontier, so you can list your capacity for resale.\n\n**What if I am a _very large_ buyer and these discounts are not enough?**\n\nWe are launching Requests for Quotes (RFQs) - as a buyer you can create a request for capacity you need and let providers bid to offer you the best price.\n\n**Team**\n\nIlia: Maths & Finance @ Imperial, Product (Wealth & Trading) @ Revolut\n\nRoman: CS + Math @ NUS, IOI gold, 2x Quant Research @ Citadel Securities, AI engineer @ Lovable\n\n**If you're a company buying tokens or a model provider, email us at [contact@touchmark.ai](mailto:contact@touchmark.ai) or sign up at[ https://app.touchmark.ai/signup](https://app.touchmark.ai/signup)**\n\n_Capacity transactions on Touchmark are limited to commercial producers and commercial end users in the wholesale capacity market. Sellers offer capacity to Buyers that purchase it to consume or use in connection with the Buyers' commercial operations. Some contracts are available only in the primary markets._", "url": "https://www.ycombinator.com/launches/Sk8-touchmark-market-for-future-inference-capacity", "votes": 26, "created_at": "2026-08-14T13:43:29.567Z" } ] }, { "id": 32254, "name": "Prodigy Research", "slug": "prodigy-research", "former_names": [ "Prodigy AI" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/a2db1ee59cf597cd664444e784ae76e1f753d775.png", "website": "https://prodigy.markets", "all_locations": "San Francisco, CA, USA", "long_description": "Prodigy is the frontier AI trading research lab. We've trained the world's best foundation model for quantitative finance. We've already achieved more than 100% returns in live trading over the course of our YC batch, doubling our capital while major indices were flat. Our founders previously worked on a trading desk at Jane Street, trained frontier AI models at Google DeepMind and shipped cutting-edge AI at Apple to billions of customers. Join us as we use AI to solve the market.", "one_liner": "Training the world's best AI model for quantitative finance.", "team_size": 2, "industry": "Fintech", "subindustry": "Fintech", "launched_at": 1786153448, "tags": [ "Reinforcement Learning", "Finance", "Investing", "Trading", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/prodigy-research", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/prodigy-research.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/prodigy-markets", "x_url": "https://x.com/realProdigyCorp", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/a2db1ee59cf597cd664444e784ae76e1f753d775.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3174989, "full_name": "Michael Wang", "title": "Founder/CEO", "founder_bio": "Founder and CEO at Prodigy Research (YC S26). Previously Senior Researcher at DeepMind, training and shipping AI models powering Google products with billions of monthly users. Before that, worked on a trading desk at Jane Street, one of the most prestigious and profitable trading firms in the world.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/michael-w-02810b161/", "x_url": "https://www.x.com/michaelwangelo", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/011ab6f302bb5ee939865ee4899d28ee9ea6acd8.jpg" }, { "user_id": 3203618, "full_name": "Yuhua Wang", "title": "Founder/CTO", "founder_bio": "Founder and CTO at Prodigy Research (YC S26). Previously shipped cutting-edge AI models and agents at Apple to billions of customers. Before that, built Salesforce's flagship AI platform serving hundreds of millions of user requests every day.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/yuhua-w-64348a66", "x_url": "https://x.com/yuhuawan", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8ab0305694f8e67bd288110245604bf4db4e3a31.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109747, "title": "Prodigy: The Frontier AI Trading Research Lab", "tagline": "Training the world's best foundation model for quantitative finance", "body": "**TL;DR** Prodigy Research is an AI lab that has trained the world's best foundation model for quantitative finance. We’ve already achieved 140% returns in live trading over the course of our YC batch, while major indices like the S&P 500 and the Nasdaq were flat or down.\n\n**Traction**\n\n![uploaded image](/media/?type=post&id=109747&key=user_uploads/3203618/c02ef51e-f0e4-43b8-a335-b55e66d5afd2)\n\nOur model beats Claude Fable and GPT-5.6 Sol at financial and quant research tasks. Prodigy's strategies are delta-neutral, and thanks to our rigorous and comprehensive risk controls, we've never had a down week.\n\n\n\n**Why Now**\n\nTrading firms generated hundreds of billions in profit last year, and this number is growing quickly YoY. Global financial markets now exceed $175 trillion in value across equities, derivatives, and digital assets. Prodigy's agents find new opportunities and alpha in the financial markets faster than any team of human quant researchers.\n\n![uploaded image](/media/?type=post&id=109747&key=user_uploads/3203618/31e0abc2-381b-4cdb-8d4f-86498b669b39)\n\n**The Team**\n\nOur founders are two brothers who left seven-figure jobs to build the future of quant trading. We bring together over 16 years of combined experience in quantitative finance, frontier model research, and large-scale AI infrastructure. Michael was a senior researcher at **Google DeepMind**, where he trained SoTA foundation models. He previously worked at **Jane Street**, one of the world’s most successful quantitative trading firms. Yuhua built cutting-edge AI at **Apple** serving **billions of users** across the world.\n\nA new era has arrived in the world of quantitative finance. To learn more, contact [founders@prodigy.markets](mailto:founders@prodigy.markets).", "url": "https://www.ycombinator.com/launches/SY7-prodigy-the-frontier-ai-trading-research-lab", "votes": 34, "created_at": "2026-08-10T18:30:00.360Z" } ] }, { "id": 32256, "name": "Zaplar", "slug": "zaplar", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c3ccfe3188152229b75f7e38241a91f40b4c9edd.png", "website": "https://zaplar.com", "all_locations": "Stockholm, Stockholm County, Sweden", "long_description": "Zaplar is the AI-native operating system for hotels.\r\n\r\nHotels run on legacy software and a fragmented tech stack. Zaplar replaces the entire hotel tech stack (PMS, POS, booking engine, etc) with one software that can handle hotel operations autonomously.\r\n\r\nLearn more at https://zaplar.com", "one_liner": "Agentic Hotel Operating System", "team_size": 7, "industry": "B2B", "subindustry": "B2B -> Operations", "launched_at": 1784823049, "tags": [ "B2B", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Operations" ], "regions": [ "Sweden", "Europe" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/zaplar", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/zaplar.json", "year_founded": 2026, "city": "Stockholm", "country": "SE", "linkedin_url": "https://www.linkedin.com/company/zaplar", "x_url": "https://x.com/ZaplarInc", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c3ccfe3188152229b75f7e38241a91f40b4c9edd.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3037509, "full_name": "Douglas Solberg", "title": "Founder/CEO", "founder_bio": "CEO of Zaplar. Previously studied physics and mathematics at the University of Chicago. At 14, I started a web development and cinematography business, with customers including some of the biggest brand in the Nordics, and has award-winning photographs. At 18, I started a nonprofit that built 30 solar systems in Tanzania. I also had a SWE internship at Newscoop and held an internship at the Becker Friedman Institute for Economics.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/douglas-solberg/", "x_url": "https://x.com/douglassolberg1", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/0da26ea47d66e82ed4773d6f3ef152036d41fc20.jpg" }, { "user_id": 3424482, "full_name": "Axel Andersson Lingbert", "title": "Founder/COO", "founder_bio": "Co-founder at Zaplar. Previously worked in consulting for banks and debt collection companies, and in hotels and restaurants, where I saw firsthand the problems Zaplar solves.\n\nMSc engineering mathematics Lund Faculty of Engineering (LTH) dropout; co-led anti-money laundering (AML) and counter-terrorist financing (CTF) group at major Swedish bank.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/axel-andersson-lingbert", "x_url": "https://x.com/axelanderssonl", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d10aa88068d9616b0640d057136e583a589ed109.jpg" }, { "user_id": 3167383, "full_name": "Oscar Karlstromer", "title": "Founder", "founder_bio": "Founder at Zaplar. Previously Honors List at Stockholm School of Economics (dropout); founder of the Nordic's largest non-profit for middle school academic competitions (40k students, 13-person team); worked APAC B2B sales, leading projects with accounts >$1M AUD across Sydney and Singapore.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/oscarkarlstromer/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a19c12ac636c64a126575a4121ca42101926093e.jpg" }, { "user_id": 3423827, "full_name": "Jon Milles", "title": "Founder/CTO", "founder_bio": "Coding since 12 MSc computer science Lund Technical University (LTH) dropout Security Software Engineer at the largest cybersecurity firm in the Nordics at 17 part of first LTH Formula Student team to achieve autonomous driving.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jon-milles", "x_url": "https://x.com/jnmlsxyz", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6265232b2a43e165fca4006adf3b0800ee95807c.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107463, "title": "Zaplar - Replacing PMS and the Hotel Tech Stack.", "tagline": "Replacing the fragmented hotel tech stack with one system that can understand and run hotel operations.", "body": "Hi everyone, we're Axel, Oscar, Jon, and Douglas, founders of Zaplar.\n\n**TL;DR:** Zaplar is an agentic hotel OS that replaces the fragmented hotel tech stack (PMS, booking engine, channel manager, POS, revenue management, housekeeping, and more) with one system built from the ground up for AI. We're eight people with backgrounds in hotel operations, engineering, AI research, and cybersecurity.\n\n![uploaded image](/media/?type=post&id=107463&key=user_uploads/3037509/c7720ef9-5a08-43eb-9631-4fefb2d239b6)\n\n**Ask:** Intros to hotel owners and operators who are unhappy with their PMS, and to executives and C-suite at hotel groups planning their next core system replacement. Details at the bottom. Reach us at [founders@zaplar.com](mailto:founders@zaplar.com).\n\n[Launch Video!](https://youtu.be/bpfJATRnYQM)\n\n## The problem\n\nIncumbents are bolting AI onto systems built for a different era. They provide a layer of AI to the end user, but the automations beneath it stay constrained. Their claim to use AI is correct. They just can't reason across the operation or automate work end to end.\n\nNo existing PMS or bolt-on solution gives hotels the architecture AI actually requires. And even if one did, a PMS alone still wouldn't automate hotel operations. It's one of as many as 15 systems a hotel runs, each built on a different architecture, each focused on one part of the operation, none able to act beyond its own boundaries.\n\nAI can't run a hotel if it can't understand the hotel.\n\nTo unlock the full potential of AI, a hotel has to run on a single system with a unified architecture: one that lets AI reason across the entire property, free of the boundaries integrations create and the context each one isolates behind them. Incumbents can't build this without replacing the foundation their business depends on, and they won't.\n\nWe arrived at this the slow way. Douglas's family has spent their careers running hotels, and we first built Zaplar to solve event bookings for them. In demo after demo, hotels were already evaluating another PMS, including ones that had just switched. Even leading vendors did not make a product hotels wanted to use.\n\n## The solution\n\nZaplar was built on this thesis from the start: the first truly AI-native operating system for hotels. It’s the path of most resistance, but the only way to do it right.\n\n![uploaded image](/media/?type=post&id=107463&key=user_uploads/3037509/5a0f8f2c-bcb0-4ebc-a3e1-e2c30c80baec)\n\nZaplar spatially maps out the entire property and its resources so that AI agents can reason across the hotel and understand how everything connects. On top of this architecture sit a PMS, POS, channel manager, booking engine, guest portal, housekeeping system, and more, with no boundaries between them. The result is an autonomous system that works in the background and completes work on its own. The system surfaces suggestions for staff to review, edit, and approve.\n\nOur goal is to be vertically integrated. Beyond deeper automation and richer context, the greatest benefit is that we can put all our effort into staying at the forefront, because we're not constrained by what anyone else is doing. Major hotel groups are starting to replace their aging systems, and we intend to be the only option on the table designed for what comes after.\n\n![uploaded image](/media/?type=post&id=107463&key=user_uploads/3037509/ad3b8b01-9f05-41f1-877e-952a35dc8e8a)\n\n## Our ask\n\n**Hotel owners and operators.** If you own or operate a hotel that is unhappy with its PMS, we would love to talk directly.\n\n**Executives and C-suite at hotel groups.** This is the intro we want most. Anyone at the VP level or above at a group, especially COOs, CIOs, CTOs, and heads of operations or commercial, and above all anyone whose group is planning a core system replacement in the next 18 months.\n\n**Everyone else.** You do not need to know their software situation to make the intro useful. We are focused on independents, small and mid-size groups, and boutique and luxury properties. If you have stayed somewhere you loved, send us the name and we will take it from there.\n\nReach us at [founders@zaplar.com](mailto:founders@zaplar.com) or learn more at [zaplar.com](http://zaplar.com).", "url": "https://www.ycombinator.com/launches/RxH-zaplar-replacing-pms-and-the-hotel-tech-stack", "votes": 20, "created_at": "2026-07-31T14:37:04.324Z" } ] }, { "id": 32257, "name": "Spectre Intelligence", "slug": "spectre-intelligence", "former_names": [ "Spectre Capital" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/dd275e68bdab3f73081986be229ec8d9f1423ae5.png", "website": "https://www.spectreintelligence.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Spectre Intelligence is a neotrading firm. We create AI traders instead of hiring human ones. Currently trading semis + bio.", "one_liner": "We are a multi-strategy trading firm that trains AI traders. ", "team_size": 2, "industry": "Fintech", "subindustry": "Fintech -> Asset Management", "launched_at": 1787421048, "tags": [ "Biotech", "Investing", "Trading", "Semiconductors" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech", "Asset Management" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/spectre-intelligence", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/spectre-intelligence.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://linkedin.com/company/spectre-intelligence-ai", "x_url": "https://x.com/spectreintellig", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/dd275e68bdab3f73081986be229ec8d9f1423ae5.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 1251785, "full_name": "Michelle Li", "title": "Founder/CEO", "founder_bio": "Co-founder and CEO at Spectre Intelligence, on leave from Harvard (I studied Economics & Biology and was originally class of '27). Previous research across generative protein language models, bacteriophages, CRISPR Cas9, and more. ISEF grand award winner, RSI and STS scholar.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/michelle-zhiya-li/", "x_url": "https://x.com/michellezli", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6eacc1c85541c9e6f7a907d66e8c4b4d53136b3b.jpg" }, { "user_id": 2574740, "full_name": "Joshua Zyzak", "title": "Founder/CTO", "founder_bio": "Founder at Spectre Capital, CS and EE at Harvard, researched deep-RL controllers for quadruped robots and multimodal LLM-robotics", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/josh-zyzak/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3a94186fa13fa1983bac60969d2c2ddafff69b21.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112001, "title": "Spectre Intelligence - a neotrading firm", "tagline": "We're an AI trading research firm, currently in semiconductors and biotech.", "body": "TL;DR: Spectre Capital is a multi-strategy trading firm run by AI agents, not human analysts. We deploy real capital to capture every decision, reasoning trace, and ground-truth outcome, so our agents learn and compound from every mistake. Most AI trading labs pay egregious amounts for training data. Ours pays us when we win.\n\nWe left Harvard after starting a hedge fund from our college dorm (go Currier!).\n\nNow we’re building Spectre Intelligence, a neotrading firm training superhuman AI traders. \n\n**The problem**\n\nIntelligence is getting cheaper, and it will rewire trading. But the scarcest input to markets isn't compute or data feeds or latency, but rather judgment. That data is fragmented, scarce, and the foundation of entire careers. It is also entirely proprietary and guarded fiercely by firms.\n\n**What we do**\n\nWe collect judgment by trading live. Every position our agents take generates decisions, full reasoning traces, and ground-truth P&L. We think this is the ultimate alignment of incentives in one of the largest TAMs in the world.\n\nThe structure is evolutionary from first principles: open competition inside each pod, and strategies and pods are born and killed constantly. \n\nUnlike a traditional pod shop, spinning up a new pod costs compute, not headcount. Our agents work superhuman hours, monitor a superhuman number of things, and learn from every real and simulated trade. Our AI traders develop specialized strategies in parallel, allowing us to scale far beyond the limits of human bandwidth. This is an edge that scales, compounds, and isn't held by any one employee.\n\n**Traction**\n\n\\- Two live pods: semiconductors and biotech equities. (We picked the verticals, not our AI traders.)\n\n**Why now?**\n\nAI is the worst it will ever be, LLM traders aren't even good yet, but we believe the writing is on the wall and want to be on the frontier. Quant funds were arguably the first agent labs: ML theory embedded for decades, quants as very expensive agents. We're building the generation after.\n\n**The ask**\n\nIf any of this resonates, feel free to reach out: michelle@spectreintelligence.ai\n\nWe’re also collecting an expert network across semis + biotech, so if you have experience, feel free to join at ", "url": "https://www.ycombinator.com/launches/T8T-spectre-intelligence-a-neotrading-firm", "votes": 5, "created_at": "2026-08-24T15:47:00.614Z" } ] }, { "id": 32259, "name": "screenpipe", "slug": "screenpipe", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/144e7207ec725036f43cc9a6b577a8320151e103.png", "website": "https://screenpipe.com", "all_locations": "San Francisco, CA, USA", "long_description": "Screenpipe records and learns how people in your company work and turns it into a searchable memory, documents, and AI agents. Model agnostic and data is stored on your device and/or where your team choose, source available.", "one_liner": "Turn your company's computer work into AI agents.", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Productivity", "launched_at": 1783386650, "tags": [ "SaaS", "B2B", "Enterprise", "Enterprise Software", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Productivity" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/screenpipe", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/screenpipe.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/screenpipe", "x_url": "https://x.com/screenpipe", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/144e7207ec725036f43cc9a6b577a8320151e103.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 716431, "full_name": "Louis Beaumont", "title": "Founder/CEO", "founder_bio": "Previously French intelligence (acquired) and hyperscale cloud observability at OVHCloud.\nBuilt video game bots business when I was 14. My training dataset: 150 books a year.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/louis030195/", "x_url": "https://x.com/louis030195", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/05f032b3236e73cfbdc174076161cd570a964593.jpg" } ], "jobs": [ { "id": 103591, "title": "Forward Deployed Engineer (Enterprise)", "url": "https://www.ycombinator.com/companies/screenpipe/jobs/EudBGQb-forward-deployed-engineer-enterprise", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$120K - $240K", "equity_range": "0.25% - 1.00%", "experience": "6+ years", "visa": "US citizen/visa only", "skills": [ "React", "Rust", "TypeScript" ] }, { "id": 109046, "title": "Senior Product Manager", "url": "https://www.ycombinator.com/companies/screenpipe/jobs/p3C1CYI-senior-product-manager", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Product", "role_type": null, "salary_range": "$160K - $220K", "equity_range": "0.25% - 1.00%", "experience": "6+ years", "visa": "US citizen/visa only", "skills": [] }, { "id": 111400, "title": "Founding Engineer", "url": "https://www.ycombinator.com/companies/screenpipe/jobs/hT5zAor-founding-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$120K - $240K", "equity_range": "0.50% - 1.50%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [ "React", "Rust", "TypeScript" ] } ], "news": [], "launches": [ { "id": 105615, "title": "Screenpipe: Record how you work and turn that into agents", "tagline": "Record how you work and turn that into agents", "body": "screenpipe records what you do, see and hear on your computer, makes it searchable, and gives agents the context to understand how you actually work. The project has 20K+ GitHub stars.\n\n\n\nMost AI agents start every task with partial context.\n\nThey can read the prompt and apps you connect. They do not remember the meeting where a decision was made, the tab you opened last Tuesday, the exact steps you took to fix a problem, or the edge case you discovered while doing the work.\n\nscreenpipe runs quietly in the background and captures screen and audio context on your machine. Then you can ask:\n\n* maintain a second brain for every person, projects, tasks, meetings\n\n\n* turn the workday into a daily recap and time breakdown for standups or client billing\n* maintain a personal CRM from the people, conversations, and follow-ups seen during the day\n* find commitments mentioned in meetings or messages that never made it into a task list\n* analyze repeated work and produce a list of workflows worth automating\n* save prompts and context from Claude, Codex, and ChatGPT into a searchable journal\n\nAgents can use screenpipe as long-term memory, so they work from what actually happened instead of only what you remembered to paste into a prompt.\n\nThe same foundation also works across team machines. Teams can map repeated workflows, generate SOPs, find bottlenecks, and identify automation opportunities across the tools where work actually happens.\n\nscreenpipe is open source and local-first. You control what gets recorded, can exclude apps, windows, and URLs, and choose where your data lives.\n\nI’ve been building my own AI second brain since 2020, but it could only remember what I explicitly saved. I built screenpipe so my AI could understand what I was actually doing across my screen and conversations. Other people began using it for the same reason. Today, we’re turning that foundation into context for personal and team agents.\n\n**Our ask**\n\nTry screenpipe and comment with the agent you would build if it could remember everything you’ve worked on.\n\nTry screenpipe: ", "url": "https://www.ycombinator.com/launches/RTT-screenpipe-record-how-you-work-and-turn-that-into-agents", "votes": 23, "created_at": "2026-07-14T14:38:00.824Z" } ] }, { "id": 32260, "name": "Manifold", "slug": "manifold-2", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/98f03122e41743439cec57f8152eac9cd506d714.png", "website": "https://www.manifoldindustries.ai/", "all_locations": "Los Angeles, CA, USA", "long_description": "The US Warehouse industry spends $75bn on labor each year. Yet, only 10% of US warehouses have any form of automation, because existing robotics are too expensive and hard to integrate. \r\n\r\nManifold deploys autonomous robots that handle picking, loading / unloading, and palletizing — priced by pick. Integration in hours. No upfront costs. No retrofitting. No need to rebuild the operation around the robot. \r\n\r\nOur mission is to make it easy for any warehouse to adopt robotics.", "one_liner": "Drop-in robotic labor that works on Day 1", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1784541343, "tags": [ "Hardware", "Warehouse Management Tech", "Robotics", "Logistics", "Supply Chain" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": true, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/manifold-2", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/manifold-2.json", "year_founded": 2026, "city": "Los Angeles", "country": "US", "linkedin_url": "https://www.linkedin.com/company/manifoldindustries", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/98f03122e41743439cec57f8152eac9cd506d714.png", "app_video_url": "https://bookface-images.s3.us-west-2.amazonaws.com/bf-public_profile/32318/app_video.mp4?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ASIAQC4NIECAOYE2QNZ5%2F20261020%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20261020T022739Z&X-Amz-Expires=60&X-Amz-Security-Token=IQoJb3JpZ2luX2VjEHEaCXVzLXdlc3QtMiJIMEYCIQDZr7mht6VpZroejzIM4ovdqqWjmh9u8AZ4wJ%2BcxZeHaQIhAN0TyP1C5dfPpTpdf3vcwz%2B044IBUbnJ55rbsOHYRWQWKuUDCDoQABoMMDA2MjAxODExMDcyIgxycpAO09Mq3A1cvyQqwgN39qg2Ox70YWArW8Z19bx1LbIV2O3f8r7xSc%2BDpl3XtmT02s9DHm%2FdyEMimOoBk5McFrR%2FThfgO9h1Zcmt1NqN1y6d1lWIsDAdluAWIqXM089dX1KVjFmsSwg8hWS2%2F4ysZFTLuD5VBUHeANJPyfmRKN9Kb4JepbPSE01eipdz4aqPxRqz%2FVUJEYLUTh5YJM6jJKNzl5l2mnWSSFiFdUfT5OS16LHDCYH4PGt9WRY2iseRxMu7FZcICWM0P8aN14WLxKsu1H0yVxk8LS%2BNbXezGhQnkMsa3bf7aMNU7x4TSs7KyM01QgiCZxC01YV43U4O8vFENPThr%2FHtKaAvtZ0%2BXSTwUE691O8mG15thNVkorInjIhilxdazy%2BzEYcGsAMNZtfx0Bd5Xn43Xf7b%2Fie4vv0d0RCHqbCGLmxRHj5VG3Pdta5Kju3%2Fpb8n%2FKmRjY%2FtuIfi9oTYAjc4PVyikHTm8UXURiHmaM0tqXZjxyh0iIponEKehSmtjmjudiXLWICFBoPhEXeaUeyK2AT6btvH4fclsQa0qRmkxdgguiWvrYxLKhNNfPLTc5QA0jlPSRoDJ%2F%2BxNHLFrGusOX7IwU6hfRMw6P3a1gY6pAETos2BWzWJY973oh8s1gy0zFzSYQP7N%2B39nIjIA7I0ymOVnRNJktoZRF6GV9DH3ymBPT2fnMgKvEaFAqVR%2F0WVwXz3ixgSwFGrvlwPr39cuFnKBEcdnivQJOXkexjqUxi9alVn76GV8ONTXD1ZupAQeLYI09gLaaAmJlyJPRAUCybD64XlECdPqS2TV2e%2FZlA5duJD1GekpDGthoc0y8PIc1ahwA%3D%3D&X-Amz-SignedHeaders=host&X-Amz-Signature=42ea8e2882267580076ab7c55512b80c4f199c59a9ac3a17a1898ff3c4ae9bc2", "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 3169788, "full_name": "Nicolas Yeh", "title": "Founder/CEO", "founder_bio": "Founder at Manifold. Grew up in a 3rd-gen supply chain family business (major consumer products supplier to US retailers). Previously at Balyasny Asset Management covering analog semiconductors ($1B+ book). Before that, Hotel Acquisitions at Starwood Capital Group, where he sourced a $60M hotel sale and was sole analyst responsible for asset management of Extended Stay America ($7.6B JV with Blackstone). BA from UPenn.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/nicolas-yeh/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2b04b9f9fd580701776cb120a1b47e822cd2f803.jpg" }, { "user_id": 3614218, "full_name": "Joshua Ibrahim", "title": "Founder/CTO", "founder_bio": "Founder at Manifold. PhD in Applied Mathematics from Caltech. Previously Software Engineering Intern at Facebook, working on Messenger P2P services and Instagram's misinformation detection systems. MS and BS in Electrical and Systems Engineering from Penn.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/joshibrahim/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2290e4ffd9df33d56490b230a84f5929381117ae.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112000, "title": "Manifold - Robotic labor for the supply chain industry", "tagline": "Robotic labor for the supply chain industry", "body": "**What if you could hire a robot the same way you hire a temp worker?** \n\nNick grew up working in his family’s 3rd-generation distribution business, where he saw firsthand how much warehouses still depend on manual labor. The obvious answer should have been robotics, but for true labor replacement, existing robots were too expensive, inflexible, and difficult to deploy. As a result, most warehouses still say they’re “not ready” for automation.\n\nNick came together with his best friend Josh, a fellow UPenn alum and PhD at Caltech specializing in safety-critical control for robotics, to solve this problem.\n\nManifold is building robotic labor for the supply chain industry. Its robots perform picking jobs from start to finish faster than humans, with no upfront cost or infrastructure changes. Customers simply pay per pick, at a cost below human labor.\n\nThey’re already deploying robots to perform more than **5 million picks across multiple warehouses.**\n\nIf you’re a warehouse operator, reach out to [nick@manifoldindustries.ai](mailto:nick@manifoldindustries.ai)\n\n", "url": "https://www.ycombinator.com/launches/T8S-manifold-robotic-labor-for-the-supply-chain-industry", "votes": 4, "created_at": "2026-08-24T18:00:00.769Z" } ] }, { "id": 32261, "name": "Rasyn", "slug": "rasyn", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ffda7ce34aee49cfec95fa8aef9ed1a70e4c3e28.png", "website": "https://www.rasyn.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "We build foundational AI models for chemistry. Almost nothing you buy is a single chemical. Products like datacenter coolant, GPU thermal paste, paint, and glue are formulations, which are mixtures of around ten ingredients each, and demand for them has grown quickly because of AI and EVs.\r\n\r\nDesigning a new formulation is essentially a search problem, because there are more possible mixtures than anyone could ever test, and none of them can be predicted in advance. As a result, bringing a single formulation to market today takes three to five years of trial and error in a lab.\r\n\r\nOur models predict how a mixture will behave far faster than the standard methods used today, which lets us design new formulations in weeks rather than years, at a fraction of the cost. To prove this, we used them to invent three chemicals that never existed before, including a PFAS-free coolant for datacenters, and we have since synthesized and patented all three.\r\n\r\nWe are already working with some of the largest manufacturers to design formulations for them, and we are building our own lab to generate the data that will train our next models.", "one_liner": "Making chemicals 100,000x faster", "team_size": 3, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1785736030, "tags": [ "Deep Learning", "Biotech", "Manufacturing", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/rasyn", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/rasyn.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/rasyn-ycs26/", "x_url": "https://x.com/rasyn_lab", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ffda7ce34aee49cfec95fa8aef9ed1a70e4c3e28.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 3475945, "full_name": "Ansh Tiwari", "title": "Founder/CEO", "founder_bio": "Building AI for scientific discovery | Physics @ Caltech, ex-NASA, Author of 3x ICML papers | Masason Foundation Member", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ansschh/", "x_url": "https://x.com/4n5che", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3629530f2959475271a8d4c059a54c800cccaed5.jpg" }, { "user_id": 972683, "full_name": "Ayush Chauhan", "title": "Founder/CTO", "founder_bio": "AI for chemical discovery | ex-Speechify (50M+ users) & 2x YC engineer | NeurIPS-published ML research, used by OpenAI & Meta", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ayushchauhan0203/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3f433dd02a059cbf14b5ec7764ee6592562cddc2.jpg" }, { "user_id": 3481804, "full_name": "Daood Hashmi", "title": "Founder", "founder_bio": "Researching in the Arnold and Marcus Lab at Caltech", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/daood-hashmi/", "x_url": "https://x.com/hash_me_NOT", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/951a6c5546701cb614078a22a24b949c93cf48a8.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108802, "title": "Rasyn: One place to run any Chemistry task", "tagline": "Towards a General Intelligence for Chemistry", "body": "**Hey!**\n\nWe're **Rasyn**, and we built a platform called **Marigold**. Give it any chemistry task in plain language and it pulls the literature, reads your files, picks the tools, sets up and runs the calculations on its own machine, and hands back the numbers along with its reasoning what it did to get them. \n\nIt's live at [app.rasyn.ai](http://app.rasyn.ai); here's a short demo: \n\n\\\n**Here's the team:** \\\nOur team has worked at NASA JPL, and researched with scientists from Nvidia and Deepmind. Our team has worked on softwares serving 50 million users, and their research was used by frontier AI labs including OpenAI and Meta. Our team studies Computational Chemistry at Caltech, has done Total Synthesis for 4+ years, and researched alongside Nobel Laureates, so they have faced these problems firsthand. They'd lose hours deciding which tool to call next, moving outputs between disconnected tools, and running a single program at a time.\n\n**Problem:** \\\nChemistry has countless software. There's one for predicting a structure, another for proposing a route, another for reading a spectrum. None of them talk to each other, so you're stuck feeding outputs between them; this clerical work is where the time goes, none of which is science.\n\n\\*\\*Solution:\\*\\*\\\nWe tried coding agents first. They're built to write software, so a chemistry task gets handled as a scripting problem and the chemistry goes unverified. Marigold checks between steps and stops when something doesn't hold.\n\nIt runs on its own machine, so it installs and executes what it needs. It works with the tools chemists already use, submits jobs to your cluster, and runs on whatever model endpoint you have, so your data can stay in the building. In the demo we hand it four thousand compounds and ask which are worth screening. It does the whole thing itself.\n\nLonger term we're building toward a general intelligence for chemistry. Marigold today reasons over existing tools, and that's where we start.\n\n**The Ask:**\n\nThis is only the first version; it will break. And we'd like it to break in front of chemists, so if you run any sort of chemistry, please try it and tell us where it falls over. \n\nWe'd also love intros to any lab doing chemistry, computational or wet, since we want to sit down and work through a real problem with people. \n\nHappy to answer any questions!", "url": "https://www.ycombinator.com/launches/SIs-rasyn-one-place-to-run-any-chemistry-task", "votes": 9, "created_at": "2026-08-03T19:40:21.411Z" } ] }, { "id": 32262, "name": "Alkera AI", "slug": "alkera-ai", "former_names": [ "Alkera AI, Inc.", "Alkera AI", "Alkera AI, Inc." ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/eecc1a49c35b3a26973c35ce53d1c3f862a058cb.png", "website": "https://alkera.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Alkera is a data engineering/analysis/science agent that works in your IDE or CLI. If you’re tired of agents that produce false, or even dangerous, reports about your data, then Alkera is for you. Our agent is equipped with your data’s history and lineage, your team’s knowledge, and tailored plugins for data apps such as Snowflake and dbt. Simply install and connect your services to begin doing trusted agentic data work. We also offer VPC & on-prem deployment options for when data can't leave your boundary.", "one_liner": "Reliable and safe data engineering and data science agents", "team_size": 3, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1783697969, "tags": [ "Developer Tools", "Data Science", "Data Engineering", "Enterprise Software", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/alkera-ai", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/alkera-ai.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/alkera-ai", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/eecc1a49c35b3a26973c35ce53d1c3f862a058cb.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 3652826, "full_name": "Rick Gao", "title": "Founder/CEO", "founder_bio": "Founder at Alkera AI, Inc. Previously at Verition Fund Management where he built financial models and optimizations for treasury financing. Previous computational biology research at Yale's Gerstein Laboratory. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rgao09", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/515662055ce684fc796921959375d09b477afbfa.jpg" }, { "user_id": 3406186, "full_name": "Tony Li", "title": "Founder/COO", "founder_bio": "Founder at Alkera AI, Inc. Previously Agentic Software Engineer at ByteDance, where he built a Python framework for multi-agent AI workflows and deployed production bots used across company teams. Former Quantitative Developer at Carthage Capital. BS in Computer Science and Mathematics from Yale.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/tonyli4722", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9e00bcde4966ff8eb6ea44e6847dce0d34f0c0b5.jpg" }, { "user_id": 3673412, "full_name": "Andrew Tran", "title": "Founder/CTO", "founder_bio": "Cofounder at Alkera AI. Prev SWE intern at Ramp and HRT, working closely with data teams. CS B.S/M.S from Yale. Prev ML research at Yale's Gerstein Lab developing biology reasoning models.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/andrewtran-cs/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fbb518656ab4ebfa5220bc5c8406b2716cf75c60.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 106204, "title": "Alkera - The data agent that you can trust", "tagline": "Alkera is an agent that replaces your coding agent to reliably and safely perform real data engineering/analysis/science work.", "body": "**TL;DR:** Data teams adopting AI agents want speed, but one wrong step can quietly ripple through your dbt models, dashboards, and business decisions before anyone notices. [Alkera](https://alkera.ai) **is an AI data agent for data engineers/analysts/scientists or anyone who wants to gather data insights that actually understands your data stack** (column-level lineage, graph-based planning, and a living knowledge base), so it ships data work you can trust. It's **#1 on UC Berkeley's [DataAgentBench](https://ucbepic.github.io/DataAgentBench/),** which tests complex, data-oriented tasks.\n\n\n\nCoding agents can write data code that runs. **For data engineering, knowing it's correct is the hard part.** This is knowing that a transformation didn't quietly redefine a column three models downstream and corrupt the revenue dashboard that the exec team checks every Monday. **For data analysis and data science, beyond data access and lineage problems, coding agents struggle with highly parallel work spread across multiple machines.** When interacting with remote computing nodes, they do so only via fragile chains of SSH commands. Long processing and training runs have you more as a babysitter than a scientist.\n\n**Alkera is a data agent that works natively across your entire data stack, with the context and guardrails coding agents lack.** Every change it makes to a query, transformation, model, or notebook is backed by a real understanding of your pipeline.\n\nIt's **#1 on UC Berkeley's DataAgentBench,** which tests complex, data-oriented tasks. You reach it through an IDE extension (VSCode, Cursor), a CLI, or the browser — wherever you already work.\n\nWhat makes it different:\n\n* **Column-level lineage across your data stack.** Trace a single row from Snowflake, through your dbt models, to your Hex dashboard. Before the agent changes anything, it maps the _blast radius_ — every downstream table and dashboard the change would touch.\n* **Graph-based planning with flexible computation.** Plan complex data tasks with parallel agents from a quick analysis to days-long modeling experiments to run across your machine and remote CPU/GPU nodes.\n* **A living knowledge base.** Alkera learns your schema docs, trusted queries, and the tribal conventions that never make it into documentation — written by both agents and humans, with humans verifying what the agent adds.\n* **Production protection.** Fine-grained permissions over every tool, including bash and SQL. Block a seemingly innocent SQL query with a dangerous UPDATE hidden away in a subexpression. Enforce OAuth to connect your warehouses to maintain existing data permissions.\n\n# **Try it**\n\nDownload Alkera at or read the docs at . For data engineering, point it at your stack and **watch it trace a change through your pipeline**, lighting up everything downstream. For data analysis and data science, have it **spin up parallel research agents** to explore your data. Reach out to [contact@alkera.ai](mailto:contact@alkera.ai) to see how we can tailor Alkera to your specific data apps/tools and computing environment.\n\n# **Our team**\n\n* Andrew Tran studied CS at Yale. At Yale, he trained LLMs to reason over multi-modal biological data. At Ramp and Hudson River Trading, he worked closely with data teams to build out ergonomic tooling for data engineers and researchers.\n* Rick Gao studied CS at Yale. He was previously at Verition Fund Management, where he worked with internal data teams and counterparties to develop ingestion pipelines and internal financial models.\n* Tony Li studied CS and Math at Yale. During his time at Carthage Capital and ByteDance, he worked with data engineering teams to build data pipelines, data analysis, and agentic tooling.\n\n# **Our ask**\n\nIf you:\n\n* perform data engineering work on production pipelines with tools like dbt and Apache Airflow\n* perform data analysis/data science work in Python/R and with tools like Jupyter notebooks\n* manage a data team – engineers, analysts, or scientists\n* want to enable non-data experts to create data insights\n\nCheck out and reach out at [contact@alkera.ai](mailto:contact@alkera.ai). We’d love to showcase a demo, learn more about your data teams’ pain points, and share more about how Alkera can help supercharge them.", "url": "https://www.ycombinator.com/launches/Rcy-alkera-the-data-agent-that-you-can-trust", "votes": 12, "created_at": "2026-07-17T18:50:58.787Z" } ] }, { "id": 32288, "name": "Callbook AI", "slug": "callbook-ai", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/a36bdaf62a231431b2534670e18c255d0347c7cd.png", "website": "https://callbook.ai", "all_locations": "San Francisco, CA, USA; Remote", "long_description": "Callbook AI is a collections agency operated by AI. We collect late stage portfolios by finding, contacting and negotiating with borrowers.", "one_liner": "AI collections agency for late stage portfolios", "team_size": 7, "industry": "B2B", "subindustry": "B2B", "launched_at": 1787023077, "tags": [ "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada", "Remote", "Fully Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/callbook-ai", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/callbook-ai.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/callbook-ai", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/a36bdaf62a231431b2534670e18c255d0347c7cd.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 2639620, "full_name": "Diego Avellaneda", "title": "Founder/CEO", "founder_bio": "CEO at Callbook AI. Applied ML researcher with two peer-reviewed publications on deep learning and embedded systems (TinyML on-device deep learning for indoor tracking; low-cost embedded sensing, Elsevier). Master's in IoT applying TinyML to real-time truck telemetry 6+ years in embedded ML. Former Senior Software Engineer at CI&T, where he built Audi's e-commerce car-purchasing platform. Second-time founder", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/diegoavellaneda/", "x_url": "https://x.com/diego_torres173", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/7281632ccf8f2d5188e03060a59b811c147401ed.jpg" }, { "user_id": 325909, "full_name": "Daniel Martinez", "title": "Founder & Product", "founder_bio": "CTO at Callbook AI. CPO at Treble (YC S19). Leading B2B products for 5+ years. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/daniel-martinez-b7a764164/", "x_url": "https://x.com/imwi90", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d5a62a492561652bcec5f1ba3ea2e50f03afdce6.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111183, "title": "Callbook AI: The AI collections agency for late stage portfolios", "tagline": "We find hard-to-reach borrowers, understand their situation, and negotiate using AI voice, email and SMS.", "body": "We are a collections agency entirely run by AI.\n\nWe have found out that the most effective way to collect a portfolio is by first doing a deep research to locate and understand each person´s situation, and then negotiate with them. \n\nAs a result, we've consistently outperformed every competing collection solutions.\n\n\n\nToday, we’re negotiating with 100K+ people every week. We’ve delivered historic recovery at every company we’ve worked with and won 100% of our head-to-head pilots against competing solutions.\n\nFounders **Diego** and **Daniel** bring experience across machine learning and B2B products. Diego has published two peer-reviewed papers on on-device deep learning and TinyML with Elsevier, has 6+ years of experience in embedded ML, and previously built Audi’s e-commerce platform at CI&T. Daniel, former CPO and founder at Treble (YC S19), has 5+ years of experience building and leading B2B products.\n\n**Our ask**\n\nIf you know a lender (bank, fintech, BNPL, retailer, insurer). We’d love an intro. sales@callbook.ai", "url": "https://www.ycombinator.com/launches/SvH-callbook-ai-the-ai-collections-agency-for-late-stage-portfolios", "votes": 25, "created_at": "2026-08-19T22:00:00.902Z" } ] }, { "id": 32319, "name": "Tarifflo Inc.", "slug": "tarifflo-inc", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/5f3acf13919479f5ac3643cb85da39ab63c49903.png", "website": "https://www.tarifflo.com", "all_locations": "Salt Lake City, UT, USA", "long_description": "Tarifflo is the infrastructure layer for modern customs - AI-powered filing, tariff recovery, and customs intelligence on one platform.\r\n\r\nFiling should be free. Legacy ABI software charges you per entry, per user, per module. Brokers charge for every transaction. Tarifflo's filing software costs nothing - file directly with CBP under companies own filer codes.\r\n\r\nWe make money when our clients do. 1 in 8 entries has an error that overpays duty - misclassification, tariff stacking, valuation, origin. Most importers never catch them. Our AI audits your entries, finds the overpayments, and helps file for your refund. \r\n\r\nUnder both is the same engine, live tariff intelligence covering every HTS code, Chapter 99 program, exclusion, and rate change - so the duty you pay going forward is as right as the refunds we recover looking back.", "one_liner": "Supply chain compliance and cost savings", "team_size": 5, "industry": "B2B", "subindustry": "B2B -> Supply Chain and Logistics", "launched_at": 1786674718, "tags": [ "B2B", "LegalTech", "Supply Chain" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Supply Chain and Logistics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/tarifflo-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/tarifflo-inc.json", "year_founded": null, "city": "Salt Lake City", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/5f3acf13919479f5ac3643cb85da39ab63c49903.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 3688899, "full_name": "Bryce Judy", "title": "Founder/CEO", "founder_bio": "Founder at Tarifflo Inc. Previously Trade Analyst at KPMG US, specializing in customs classification and international tax. Licensed customs broker (CBP), data scientist", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/brycejudy/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c1d92eae9ff38990e80f1d0c8ebaac4fbc3c4597.jpg" } ], "jobs": [], "news": [], "launches": [] }, { "id": 32418, "name": "Traceforce", "slug": "traceforce", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/8a31e3b43130c5974d52608c066ff45cc746112f.png", "website": "https://www.traceforce.ai", "all_locations": "San Francisco, CA, USA", "long_description": "AI apps such as ChatGPT and Claude have become part of everyone's daily workflow, but they introduce new security risks that traditional security tools weren't built to handle. Security teams at companies from startups to Fortune 500s, are struggling to keep up.\r\n\r\nTraceforce gives a company’s security team visibility of how AI apps are being used across company’s devices, and the ability to detect and prevent security breaches and unsafe actions as early as possible.", "one_liner": "CrowdStrike for AI Agents", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Security", "launched_at": 1785015608, "tags": [ "SaaS", "Security", "Cybersecurity", "Enterprise", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Security" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/traceforce", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/traceforce.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/traceforce-security", "x_url": "", "fb_url": null, "cb_url": "Traceforce, Inc", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/8a31e3b43130c5974d52608c066ff45cc746112f.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 2775802, "full_name": "Xia Hua", "title": "Founder/CEO", "founder_bio": "Founder at Traceforce. Previously Director of Engineering at Clumio, managing all flagship ransomware protection products. PhD in Applied Math from MIT.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/xia-hua-ph-d-28856a30/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2f3e3c63d811c56302968adf9fea80553cf68f79.jpg" }, { "user_id": 3675843, "full_name": "Varun Wadhwa", "title": "Founder/CTO", "founder_bio": "Co-founder & CTO at TraceForce. Previously Senior Software Engineer at LinkedIn, leading high-performance database systems for RAG. Before that, Staff Software Engineer at Clumio, leading distributed data infrastructure and backup. Earlier built distributed payments systems at Microsoft. BS in EECS, UC Berkeley.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/varunwadhwarcds", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3bb82e62b797bba9d01efa60a14cb9d26fe41a26.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 117100, "title": "Traceforce TraceGraph – Ask anything about your company's AI usage", "tagline": "Know where every AI token goes", "body": "**TL;DR:** TraceGraph turns every AI interaction in your company into a fully queryable graph linking users, apps, models, MCPs, and tokens. You can ask questions in plain English, slice the data any way you want, and build agentic workflows on top of it out of the box. Try it free at [traceforce.ai](http://www.traceforce.ai).\n\n**👋 Hi YC, we're Xia and Varun**\n\nWe're the founders of Traceforce (S26), and we're building security and visibility for AI agents running on company devices. We previously built one of the world’s most scalable cyber-resillience systems at Clumio.\n\n**🔥 The problem**\n\nAI usage is exploding, but the data about it is stuck in billing totals and static dashboards. When you want to know which team's agents are burning tokens, which MCPs get called most, or who is using AI in ways that don't fit policy, you end up exporting CSVs or waiting on someone to build a report.\n\n**🕸️ How TraceGraph works**\n\nTraceforce runs on the endpoint, so it captures AI activity across every app and agent. TraceGraph connects that activity into one graph, and that graph lets you do three things:\n\n* **Query anything.** Every user, team, app, model, MCP, and token is a connected node, so you can slice the data in any combination with no predefined dashboards.\n* **Ask in natural language.** For example: _\"Which teams' token usage doubled this month?\"_ or _\"Show Claude Code sessions that called the Supabase MCP.\"_\n* **Build agentic workflows out of the box.** Put agents on top of the graph to flag spend spikes, send weekly usage reports, or escalate risky patterns automatically.\n\n🎥 **Demo:**\n\n\n\n**🙏 Our asks**\n\n* **Try it:** Install in under 5 minutes and see your team's token usage today.\n* **Intros:** Heads of engineering, IT, finance, or security leaders managing AI rollouts and spend.\n* **Feedback:** What would you ask your AI usage data? We'll show you the answer.\n\n📩 xia@traceforce.ai\n\nBook a time ", "url": "https://www.ycombinator.com/launches/USi-traceforce-tracegraph-ask-anything-about-your-company-s-ai-usage", "votes": 7, "created_at": "2026-09-28T14:55:58.682Z" }, { "id": 107460, "title": "Traceforce — Catch risky agent actions your security gateway will never spot", "tagline": "Company-wide security monitoring for AI apps", "body": "# \\*\\*\\*\\*\n\n**TL;DR:** AI agents are quietly pulling secrets off your employees' laptops — and your security gateway can't see it. Traceforce installs in under 5 minutes and gives you a full inventory of every AI app, MCP, and skill running on every device, then lets you block the dangerous stuff in real time. Built by the ex-Clumio engineering team. Traceforce is already deployed on 6,000+ devices across 10+ enterprise customers, with 40+ active pilots worth $3M+ in ARR. Try it free at [www.traceforce.ai](http://www.traceforce.ai).\\\n\\\n**The incident that started this**\\\nAt our last company, an AI coding agent running on a developer's laptop scraped what looked like database credentials and pushed them straight into a public GitHub repo. It took us _weeks_ to investigate, rotate credentials, and clean up the fallout. We got lucky — they were dev-environment creds, so no customer was affected. Next time, nobody will be that lucky.\\\n\\\nThat's the whole problem with AI at work now: ChatGPT, Claude, Cursor, and Claude Code are everywhere on employee devices, wired into your data through MCPs and skills that no one is tracking. The tools your security team already owns — gateways, proxies, enterprise APIs — never see any of it, because the risky action happens _on the device_, before it ever hits the network.\\\n\\\n**Why us**\\\nWe're Xia and Varun. We ran engineering at Clumio, a cyber-resilience company acquired by Commvault in 2024, where Xia was Director of Engineering and Varun was tech lead. We've spent our careers building security infrastructure — and we validated this problem with 50+ CISOs and CIOs before writing a line of code. Every one of them said the same thing: \\_I need to see and control this without slowing my engineers down.\\_\\\n\\\n**Why on-device — and why gateways can't do this**\\\nCompetitors sit at the gateway or plug into enterprise APIs. That misses the self-adopted AI tools employees install themselves, and it can't see how an agent is actually connected to your data via MCPs. Traceforce runs on the endpoint, so we see the full picture — which AI apps are running, what MCPs and skills they're wired to, what data they can reach — and we can _warn or block_ an unsafe action the moment it happens, not after it's already on the network.\\\n\\\n**What you get**\n\n* Install in <5 min. Full inventory of every AI app, MCP, and skill across your devices within 30 minutes.\n* Discover shadow AI and see exactly how each agent connects to your data sources.\n* Real-time controls: warn and block unsafe actions (like uploading a client list to a chatbot) before they happen.\n* Open-source MCP pentesting: we also built [mcp-xray](https://github.com/traceforce/mcp-xray) to dynamically detect vulnerable MCPs — try it today.\n\n\\\n**Traction**\n\n* Securing 6000+ employee devices across 10+ enterprises.\n* Customer quote from a leading database provider with over 500 employees: “To govern AI, we need full visibility into AI usage before we can place any type of controls. Traceforce gives us both visibility and control we needed”.\n* 40+ Active pilots including multiple Fortunate 500 enterprises.\n\n\\\n\\\n**Our ask**\n\n* **Intros** to security engineers, security architects, DevSecOps teams, or CISOs securing AI adoption. (We're the ex-Clumio team — that's usually enough to get a conversation.)\n* **Try Traceforce free** on your own laptop — <5 min to install, and we want candid feedback, good or bad.\n* **Challenge us:** tell us which AI apps, MCPs, or workflows we should support next.\n* **Share** with anyone wrestling with security over ChatGPT, Claude, Cursor, Claude Code, or custom MCPs.", "url": "https://www.ycombinator.com/launches/RxE-traceforce-catch-risky-agent-actions-your-security-gateway-will-never-spot", "votes": 7, "created_at": "2026-07-28T16:50:42.184Z" } ] }, { "id": 32484, "name": "EdotEnv", "slug": "edotenv", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/a53586c9aacc3d5c0003f5bde0beb3535a3613f6.png", "website": "https://edotenv.com", "all_locations": "San Francisco, CA, USA", "long_description": "A Quant Neolab building toward RSI by leveraging markets as self-improving research environments.\r\n\r\nThey work with frontier AI labs and academic groups building research harnesses, evaluation benchmarks, and post-training environments.\r\n", "one_liner": "Quant Neolab building self-improving agents from quant data", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1786503507, "tags": [ "Reinforcement Learning", "Time Series" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/edotenv", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/edotenv.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/a53586c9aacc3d5c0003f5bde0beb3535a3613f6.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 2622651, "full_name": "Rui Wang", "title": "Founder/CEO", "founder_bio": "I am Rui (\"ray\" like the fish). CEO of EdotEnv, a quant neolab teaching LLMs robust research and good research taste. We use markets as an infinite source of increasingly hard data to build environments, harnesses, and post-train models.\nBackground in quant at G-Research, prev. applied maths at ETH Zurich.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rui-wang-456032198/", "x_url": "https://x.com/RRicefan", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c0681ade5300e908a4d3267d8ead62de1be2484e.jpg" }, { "user_id": 2995240, "full_name": "Michael Zhang", "title": "Founder/CTO", "founder_bio": "CTO at EdotEnv, a quant neolab building toward RSI.\n\nPreviously@Etched: Designed LLM Inference System for SOHU chips\nPreviously@TransmarketGroup: Built automated high-frequency option trading strategies\n\nETH Zürich Alumni", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/michael-zhang-845204232/", "x_url": "https://x.com/Michaelz1914", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/01fb39f6e87e8fc6884b4a54b36981b4011bb42a.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110398, "title": "EdotEnv: Quant Neolab building towards RSI", "tagline": "We turn markets into self-improving research environments for AI agents.", "body": "EdotEnv (YC S26) is a Quant Neolab building toward RSI. They turn markets into self-improving research environments for AI agents.\n\nThe path to RSI is a continuous loop: agents conduct experiments, learn from the results, and autonomously become better at conducting the next round of research. For that loop to continue, each round must present a harder problem than the last.\n\nMarkets create this dynamic naturally. Every discovered inefficiency attracts competition, and every successful strategy makes future opportunities harder to find. Markets therefore provide an advancing frontier for self-improving agents.\n\nEdotEnv captures this process in multi-step environments where agents form hypotheses, design experiments, verify results and iterate. They use these environments to evaluate and post-train agents on increasingly difficult research tasks.\n\nThey work with frontier AI labs and academic groups building research harnesses, evaluation benchmarks, and post-training environments.\n\nEdotEnv's team previously worked in quantitative research at G-Research, high frequency options trading at TransMarket Group, and LLM inference research at Etched.\n\nIf you're working on RSI, RL, post-training, or AI for research, reach out to them at [info@edotenv.com](mailto:info@edotenv.com).\n\n![uploaded image](/media/?type=post&id=110398&key=user_uploads/89898/46b2eec2-b3c9-4467-b638-d9a51b2c39cd)\n\n", "url": "https://www.ycombinator.com/launches/Sic-edotenv-quant-neolab-building-towards-rsi", "votes": 5, "created_at": "2026-08-13T21:30:00.360Z" } ] }, { "id": 32517, "name": "Standard Machines", "slug": "standard-machines", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/20aa62fc0a930e7bfdba7047f185bab421afd2b2.png", "website": "https://standardmachines.com", "all_locations": "San Francisco, CA, USA", "long_description": "We build environments and evaluations to train and evaluate models on advanced chip design.", "one_liner": "Teaching AI to Design Chips", "team_size": null, "industry": "B2B", "subindustry": "B2B", "launched_at": 1785623138, "tags": [ "Semiconductors", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/standard-machines", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/standard-machines.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/standardmachines/", "x_url": "https://x.com/stanmachines", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/20aa62fc0a930e7bfdba7047f185bab421afd2b2.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 1978488, "full_name": "Jacob Peake", "title": "Founder/CEO", "founder_bio": "Founder at Standard Machines. \nEx GPU Architect at Apple.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jacob-peake/", "x_url": "https://x.com/jacobpeake", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1e58da36d411b29af7b6af0bb04427b019ed414f.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111880, "title": "Standard Machines. Teaching AI to Design Advanced Chips.", "tagline": "Standard Machines builds environments to train and evaluate models on advanced chip design tasks. We work with frontier labs to push their model capabilities.", "body": "**The Idea**\n\nAdvancement in AI relies on chips. Transformers won because they mapped well to GPUs. Frontier Labs are now co-designing their own silicon to maximize intelligence/joule.\n\nIt takes years and 100s of experts to bring a chip from architecture through tape-out.\n\nOur goal as a company is to push the frontier of AI on chip design tasks - enabling advanced chips to be taped-out by small teams in months.\n\nWe also believe this is the path forward to achieving true recursive self-improvement (RSI). Where AI designs better chips and better chips train more intelligent AI.\n\nTo achieve this, we're building **RL Environments For Long-Horizon Chip Design.**\n\nOur environments enable 1000s of parallel rollouts for training on chip design tasks. We also curate a private held-out test set for performing evals in the environment to compare model capabilities.\n\n![uploaded image](/media/?type=post&id=111880&key=user_uploads/1978488/da8f5344-b2f0-4613-b8ef-5ee18a23300c)\n\n---\n\n**What Exists**\n\nChip design should be a perfect domain for this type of RL. Instead:\n\n\\- The academic benchmarks are one-shot puzzles of \\~100 lines, and they're in every training set.\n\n\\- The field's de facto standard (NVIDIA's CVDP) has saturated - going from 34% to 97% once agents could iterate against the grader.\n\n\\- The published tasks are short-horizon and graded by running testbenches. In one study, designs that passed their bundled testbenches 95-97% of the time were only \\~20% correct under formal checking.\n\nAnd none of it resembles the real loop: months of work, tools in the loop, and hard trade-offs between latency, throughput, area, and power.\n\n---\n\n**What We Build**\n\nOur RL Environment is designed for long-horizon chip design tasks.\n\n\\- **Long-horizon.** One episode is the real loop: choose a microarchitecture, write the RTL, simulate, synthesize, check timing and power, revise.\n\n\\- **Reliable grading.** Every task is an executable contract with exact checking. No LLM judge anywhere in the reward path. Correctness is a hard gate: a fast, small, wrong chip earns zero.\n\n\\- **Physics as the yardstick.** Every task ships with a computed speed-of-light bound: the best the physics of the task allows. Scores are reported as a fraction of that bound, so they're comparable across tasks.\n\n\\- **The full trade-off.** Real chips are trade-offs. We record area, power, latency, and throughput as an unreduced vector - labs can scalarize this into a terminal reward for training however they want.\n\n\\- **Deterministic.** Every graded result replays bit-identically on another machine.\n\n---\n\n**Why Now**\n\nChip Design is increasingly something the frontier labs care about. They are all designing their own custom inference chips. Anthropic is already hiring chip-design RL engineers at $500-850k to build these types of environments in-house - and there’s currently no yardstick that hasn’t been saturated.\n\n---\n\n**Who We Are**\n\nI was the first intern & youngest hire into GPU Architecture at Apple, working with industry-leading architects.\n\nI've seen how leading-edge chips get designed end-to-end: what the loop looks like, and how the power/perf/area trade-offs actually get graded.\n\nOur environments run that same loop & formalize that grading.\n\n**Asks**\n\nIf you’re working on RL, post-training, or data at a lab or neolab, or interested in post-training models for chip-design capabilities. Please reach out, [founders@standardmachines.com](https://www.standardmachines.com/launch).\n\n---\n\n---\n\n---\n\nRead our full launch post, [here](https://standardmachines.com/launch).\n\n\\- [Jacob](https://x.com/jacobpeake)", "url": "https://www.ycombinator.com/launches/T6W-standard-machines-teaching-ai-to-design-advanced-chips", "votes": 22, "created_at": "2026-08-24T17:21:03.769Z" } ] }, { "id": 32518, "name": "Agency Tool Company", "slug": "agency-tool-company", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/244ddf9b0965c4c601170d71dd660ad1a89f8792.png", "website": "https://agencytool.com", "all_locations": "Denver, CO, USA", "long_description": "Agency Tool Company ships software to robots in the field 20x faster than a Docker pull with ATC Deploy, sending only the bytes that changed and resuming when the network drops. Deployment tooling is the connection between the lab & the field that continually bottlenecks robotics scaling and lacks a strong robotics-native solution. Deploy is live today with 3 launch partners across ag, construction, and logistics robotics.\r\n\r\nOur second product, ATC Build, offers embedded-compute as a CI service so devs can run tests and validate their software on the actual target hardware. Build supports the Nvidia Jetson lineup, Raspberry Pi, and a number of other embedded platform targets.\r\n\r\nJack Morrison and Davis Foster founded and sold Scythe Robotics after deploying hundreds of autonomous mowers across the country; Agency Tool Company is building the robotics deployment platform across hardware and software they wish they had.\r\n\r\nLearn more at https://agencytool.com", "one_liner": "Reliable infrastructure to build, deploy and manage real-world robots.", "team_size": 3, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1784914207, "tags": [ "Developer Tools", "Robotics", "Automation", "Infrastructure", "Self-Driving Vehicles" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/agency-tool-company", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/agency-tool-company.json", "year_founded": 2026, "city": "Denver", "country": "US", "linkedin_url": "https://www.linkedin.com/company/agency-tool-company", "x_url": "https://x.com/agencytoolco", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/244ddf9b0965c4c601170d71dd660ad1a89f8792.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 3668961, "full_name": "Jack Morrison", "title": "Founder/CEO", "founder_bio": "Founder and CEO at Agency Tool Company. Previously founder/CEO at Scythe Robotics, building autonomous electric outdoor equipment from 2018-2026 and successfully exiting to ASI. Earlier co-founded Replica Labs (mobile 3D scanning) and worked as a computer vision engineer at Occipital on the Structure Sensor. BA Computer Science from Bowdoin.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/morrisonjack/", "x_url": "https://x.com/jackmorrison", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/13172ea12c502ccffb1824b2ef8e356abeabdfae.jpg" }, { "user_id": 253511, "full_name": "Davis Foster", "title": "Founder", "founder_bio": "Founder at Agency Tool Company - making it easier to scale robots. Previously co-founded Scythe Robotics ('18-26), made autonomous commercial mowers, acquired by ASI in 2026. Built and shipped hundreds of robots still operating across the country. Thiel Fellow.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/davisfoster", "x_url": "https://x.com/davistfoster", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3b4566280f9b3ba7c4ea23beb319e78fbf6db999.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107934, "title": "Agency Tool Company: The build and deployment platform for robotics", "tagline": "Field Robots Faster with ATC Deploy", "body": "Today we're sharing Agency Tool Company and our first product, ATC Deploy: a faster, more reliable way to ship software to robots wherever they are. It moves updates up to 20 times faster than docker pull or scp and works with any software stack already in use.\n\n\n\n## The updates that took until 2 a.m.\n\nDavis and I spent eight years building Scythe Robotics and putting hundreds of autonomous mowers to work across the country. For years, we deployed software with a tool I'd cobbled together for our first few robots. As the fleet grew, deployment became an ordeal: engineers stayed up until 2 a.m., we coordinated with customers to move robots out of sheds and into LTE coverage at set times, and failed transfers restarted from zero. Shipping a fix could take more coordination than building it. We had models to train and customers to serve, so the tooling limped along with incremental bandaids.\n\nWe've since heard the same story from robotics teams across industries. That deployment becomes a hidden bottleneck: teams become release-averse, feedback loops slow and improvements stall. Stitched-together web and IoT tools barely work if you know which scripts to run and who to call, so they survive, even if it means every update is distracting from improving the product.\n\n## What we built\n\nATC Deploy ships whatever you already run: OS images, Docker images, or any file tree. Run “atc ota push user@robot”, and it sends only the changed bytes, then restarts your software through hooks you define—way faster than a docker push.\n\nInstall takes one “curl | sh” command and a config file so you can get back to work fast. Interrupted transfers resume where they stopped, and an offline robot can receive updates by USB if you need. There's no new runtime, packaging format, or operating system to adopt.\n\n## Shipping to robots today\n\nATC Deploy already ships software to robots at Burro, which has deployed more than 750 machines worldwide, Tempo, polishing concrete across several US states and Canada, and Gather, who is scaling their 99.9% accurate inventory drones nationwide. Their teams spend less time managing delivery and more time improving the robot.\n\n**Operate a robot fleet? Join the ATC Deploy waitlist at [agencytool.com](https://agencytool.com/).** Know a team that should be using it? Send them this post, or to [mailbox@agencytool.com](mailto:mailbox@agencytool.com).\n\nIt's time to build robots.\n\n![uploaded image](/media/?type=post&id=107934&key=user_uploads/253511/2f53d874-b652-4819-b734-bd52256958d8)\n\n", "url": "https://www.ycombinator.com/launches/S4s-agency-tool-company-the-build-and-deployment-platform-for-robotics", "votes": 6, "created_at": "2026-07-28T15:54:16.451Z" } ] }, { "id": 32520, "name": "Snowstone", "slug": "snowstone", "former_names": [ "Cerenovus" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/5fd1252bd802d29775f848bb916f4456301ae679.png", "website": "https://snowstone.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Snowstone helps private equity firms drive operational value creation with AI. It's like a performance improvement engagement from a top consulting firm, but built from the company's own data, delivered in one week, and with every finding traced to its source. Today, private equity firms rely on consultants to find where portfolio companies can improve. That work depends on interviews and data requests, so the conclusions are secondhand, biased by internal politics and hard to check. Snowstone goes straight to the source data. It analyzes how a portfolio company actually operates and identifies automation potential, duplicated work, process bottlenecks and key-person risk, each quantified by its EBITDA impact.", "one_liner": "AI-Driven Value Creation and Margin Expansion for Private Equity", "team_size": null, "industry": "B2B", "subindustry": "B2B -> Operations", "launched_at": 1780900630, "tags": [ "Finance", "Operations", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Operations" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/snowstone", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/snowstone.json", "year_founded": null, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/snowstone-ai", "x_url": "https://x.com/Cerenovusai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/5fd1252bd802d29775f848bb916f4456301ae679.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [], "jobs": [], "news": [], "launches": [] }, { "id": 32521, "name": "Molagri", "slug": "molagri", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/71eba8e2b8b0a262e4c76ef41e5ab46b59552e6d.png", "website": "https://molagri.com", "all_locations": "Toronto, ON, Canada", "long_description": "Molagri develops safer, more effective pesticides for agriculture.\r\n\r\nWe design biopesticides that hit a precise molecular target unique to pest insects, knocking out pests while sparing pollinators, beneficial insects, and the wider ecosystem. By modeling how pests are likely to evolve, we stay a step ahead of resistance, keeping our products effective against the pests of today and tomorrow. The higher efficacy and biobased production will also contribute to emissions reductions in both manufacturing and field use.\r\n\r\nOur founders are PhD lab mates from the University of Toronto who have worked together for six years and published in Nature Communications and Nature Microbiology. \r\n", "one_liner": "Resistance free pesticides.", "team_size": 4, "industry": "Industrials", "subindustry": "Industrials -> Agriculture", "launched_at": 1782770010, "tags": [ "Synthetic Biology", "Biotech", "Sustainability", "Climate", "Agriculture" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Agriculture" ], "regions": [ "Canada", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/molagri", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/molagri.json", "year_founded": 2026, "city": "Toronto", "country": "CA", "linkedin_url": "https://www.linkedin.com/company/molagri", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/71eba8e2b8b0a262e4c76ef41e5ab46b59552e6d.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 2273508, "full_name": "Zaky Hassan", "title": "Founder/CEO", "founder_bio": "Zaky is a protein biochemist and viral structural biologist who has characterized mechanisms of antibody neutralization in human pathogenic coronaviruses. He has an undergraduate degree in Biochemistry (Honours BSc) and graduate training in Biochemistry (PhD candidate) from the University of Toronto. He was awarded $130K+ in fellowships, and published twice in Nature Communications. Zaky previously founded a non-profit, raised $1M+ in grants and operated programming for 1000+ youth in Toronto. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/zaky-hassan/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a0a2d10fb8eb3af2c5ae7e66b5bcd91e9c0fa244.jpg" }, { "user_id": 3659967, "full_name": "Min Jin", "title": "Founder/CSO", "founder_bio": "Min is a viral structural biologist and protein biochemist, and the current world record holder for the highest resolution structure of a coronavirus spike protein in the Protein Data Bank (1.68 Å). She has an undergraduate degree in Biophysics (Honours BSc) and graduate training in Biochemistry (PhD candidate) from the University of Toronto and she has published in Nature Microbiology and Nature Communications. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/min-jin-ca/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/42b06f5f78332c243bd0f7ed0a6865ddf972d2be.jpg" } ], "jobs": [ { "id": 114656, "title": "Scientist / Senior Scientist, Synthetic Virology for RNA Delivery", "url": "https://www.ycombinator.com/companies/molagri/jobs/HdOJVZM-scientist-senior-scientist-synthetic-virology-for-rna-delivery", "location": "Toronto, ON, CA", "type": "Full-time", "role": "Science", "role_type": "Biotechnology", "salary_range": "$130K - $160K CAD", "equity_range": "0.50% - 1.00%", "experience": "3+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 114689, "title": "Scientist / Senior Scientist, Insect Biology and Bioassays", "url": "https://www.ycombinator.com/companies/molagri/jobs/zvbnS35-scientist-senior-scientist-insect-biology-and-bioassays", "location": "Toronto, ON, CA", "type": "Full-time", "role": "Science", "role_type": "Biotechnology", "salary_range": "$130K - $160K CAD", "equity_range": "0.50% - 1.00%", "experience": "3+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 114690, "title": "Sr. Research Associate, Yeast Expression and VLP Production", "url": "https://www.ycombinator.com/companies/molagri/jobs/YZjWSPT-sr-research-associate-yeast-expression-and-vlp-production", "location": "Toronto, ON, CA", "type": "Full-time", "role": "Science", "role_type": "Biotechnology", "salary_range": "$85K - $105K CAD", "equity_range": "0.25% - 0.50%", "experience": "3+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 116349, "title": "Scientist / Senior Scientist, Molecular Display and Binder Discovery", "url": "https://www.ycombinator.com/companies/molagri/jobs/qHL3N9N-scientist-senior-scientist-molecular-display-and-binder-discovery", "location": "Toronto, ON, CA", "type": "Full-time", "role": "Science", "role_type": "Biology", "salary_range": "$130K - $160K CAD", "equity_range": "0.50% - 1.00%", "experience": "3+ years", "visa": "US citizenship/visa not required", "skills": [] } ], "news": [ { "title": "Molagri raises $5.8m to build ‘updatable’ biopesticides", "url": "https://agfundernews.com/molagri-raises-5-8m-to-build-updatable-biopesticides", "date": "Oct 14, 2026" }, { "title": "Molagri wants to design a pesticide that evolves as quickly as pests | BetaKit", "url": "https://betakit.com/molagri-wants-to-design-a-pesticide-that-evolves-as-quickly-as-pests/", "date": "Sep 04, 2026" } ], "launches": [ { "id": 108796, "title": "Molagri: Resistance-free pesticides 🪲", "tagline": "Species-specific biopesticides built from engineered insect viruses, updatable the moment pests develop resistance.", "body": "**TLDR:** We build next-generation pesticides from engineered insect virus-like particles. They target a single pest species, stay safe for humans, pollinators, and the environment, and can be reprogrammed over time as pests evolve resistance.\n\nPesticides underpin modern agriculture and food security, accounting for $80B in annual spend. Without effective pesticides, we can't grow enough food.\n\n\n\n**The Problem 🌱**\n\nPests are developing resistance faster than new pesticides can be invented to replace the failing ones. At the same time, regulators are banning core active ingredients over environmental and human toxicity. Farmers are losing tools from both ends at once.\n\n**Our Solution 🪲**\n\nWe build pesticides out of insect virus-like particles. We start with a virus that only infects insects and is harmless to humans. We strip out its genome so it can't infect or replicate, and we keep the empty protein shell as a delivery vehicle.\n\nWe engineer that shell with two parts:\n\n* A **targeting element** that recognizes a single pest species (or a set of species). This is what keeps the product species-specific and invisible to humans, pollinators, and beneficial insects.\n* A **payload** that kills the pest.\n\nBoth parts are swappable. Changing the targeting element will modify which pest it kills. Changing the payload will result in a new kill mechanism for the same pest, which is how we update the product when resistance appears.\n\nDesigning for true pollinator safety is hard, so we did the underlying science ourselves. We solved a cryo-EM structure of a key receptor in honey bees, which lets us deliberately engineer our targeting elements so they do not bind pollinators.\n\n![uploaded image](/media/?type=post&id=108796&key=user_uploads/2273508/62f4362a-1938-4843-ab26-c9a62c742a84)\n\n**Why Us? 🧑🏽‍🔬**\n\nWe are Zaky and Min, viral structural biology PhD candidates and lab mates at the University of Toronto, where we have worked together for the past 6 years studying how viruses interact with their hosts. We have published in Nature Microbiology and Nature Communications, and Min holds the world record for the highest-resolution structure of a coronavirus spike protein.\n\n**Traction 📈**\n\nWe have signed our first grower LOI, have more in the pipeline, and produced our first set of viruses for testing. During the batch we also solved a cryo-EM structure of a honey-bee receptor that directly informs our pollinator-safe design.\n\n**Connect with us!**\n\n* **Growers and producers**: if you are fighting resistant pests or losing tools to bans, we would love to talk about your crops and target pests. Reach us at [molagri.com](http://molagri.com) \n* **Everyone else:** intros to ag and crop-science operators, or anyone who has built regulated biological products, are always welcome.", "url": "https://www.ycombinator.com/launches/SIm-molagri-resistance-free-pesticides", "votes": 15, "created_at": "2026-08-03T18:35:06.582Z" } ] }, { "id": 32522, "name": "Peer", "slug": "peer", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/0047fd6626342e5850c5136fd718899a2c8c81cb.png", "website": "https://peer-freight.com", "all_locations": "San Francisco, CA, USA", "long_description": "Peer is building an AI-native freight brokerage that quotes freight, finds the right trucking companies, tracks shipments, and handles problems from pickup through delivery for shippers.", "one_liner": "AI-native freight brokerage", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Supply Chain and Logistics", "launched_at": 1786336832, "tags": [ "Digital Freight Brokerage", "Logistics" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Supply Chain and Logistics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/peer", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/peer.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/0047fd6626342e5850c5136fd718899a2c8c81cb.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 2454208, "full_name": "Aaron Zhang", "title": "Founder/CEO", "founder_bio": "Co-founder and CEO of Peer. Majored in Computer Science at Harvey Mudd College and dropped out for YC S26. Built AI voice agents for trucking fleets at Waylens (a trucking telematics company) and researched AI agents and computer vision at the MIT Media Lab (City Science Lab, Taipei). Led the Claremont Accelerator, which raised $124K to back student startups. 1st @ Google DeepMind Hackathon at UCLA with Felix.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aaronttzhang/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/114b28e90ca8a22744bc307f0aee643e3b422fdc.jpg" }, { "user_id": 2597858, "full_name": "Felix Peng", "title": "Founder/CTO", "founder_bio": "Cofounder of Peer. CS/Engineering at Harvey Mudd. Engineering with Apple, IBM, Magnefy, and others. At LADWP shipped internal AI tools to 250+ staff. 1st Place National, DOE EnergyTech University Prize ($25K, youngest winner). Won the Google DeepMind Hackathon with Aaron Zhang. President and founding team of the Claremont Accelerator; raised $124K from alumni, AWS, and Google to back student startups.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/felixpeng24/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/24129fee3f517ec726754cf95edd440bdbb62c6c.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110495, "title": "Peer: AI-Native Freight Brokerage", "tagline": "Peer brokers every load autonomously, from quote to delivery.", "body": "Freight brokerages in 2026 look like stock trading floors in 1989. Rooms of people on phones, quoting prices from memory, faxing paperwork across the country. Then electronic trading showed up and those floors emptied out. Peer (YC S26) is that moment for trucking.\\\n\\\n[https://drive.google.com/file/d/1IsbGimgbdogUQdADAzDloIuPn9zQgEJi](https://drive.google.com/file/d/1IsbGimgbdogUQdADAzDloIuPn9zQgEJi/view?usp=sharing)\n\nTrucking is a $900 billion industry, and 91.5% of American trucking companies run ten trucks or fewer. Nobody shipping freight can manage hundreds of those relationships, so they hand the load to a broker, who finds the right truck, in the right place, at the right price.\n\nPeer is an AI-native freight brokerage. A shipper sends Peer a load and agents do everything after that: price it, find and vet the trucking company, handle the paperwork, track the delivery, and fix what goes wrong on the road. Even the most tech-forward incumbents need thousands of people to run a brokerage. Peer runs on agents, doing far more loads with far fewer people.\n\nShippers want a fair price fast, a truck that is insured, and a trucking company that shows up. An agent quotes market rates in minutes instead of days of emails, because it reads the request the moment it lands. It catches a lapsed insurance policy before the freight moves, because it rechecks continuously, not once at signup. And it drops a trucking company that keeps failing, because it has no relationship to protect.\n\nAaron and Felix are both technical and both dropped out of Harvey Mudd to build Peer. Aaron built voice AI for trucking fleets at Waylens and researched AI agents at the MIT Media Lab. Felix worked in machine learning at Magnefy and shipped AI tools to the LA Department of Water and Power. Together they won the Google DeepMind Hackathon at UCLA.", "url": "https://www.ycombinator.com/launches/SkB-peer-ai-native-freight-brokerage", "votes": 20, "created_at": "2026-08-14T19:32:27.384Z" } ] }, { "id": 32523, "name": "Elite", "slug": "elite", "former_names": [ "Stratum", "Stratum Industries" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/950a2e88427ff9cd0ec85f5da8c2e8bd7d8333df.png", "website": "https://elitehq.ai/", "all_locations": "New York City, NY, USA", "long_description": "Elite builds AI voice agents that run the front office for healthcare practices. Elite answers every patient call instantly and finishes the job inside the practice's EHR: scheduling, triage, referrals, refills, eligibility, intake, and billing, across voice, SMS, and web, around the clock.\r\n\r\nFront offices are drowning in phone volume, so patients wait on hold, calls go unanswered, and staff get pulled away from care. Elite picks up every call, resolves the routine ones end to end, and escalates anything complex to the team with full context. It ships knowing the scheduling protocols of 18 specialties and writes back to more than 19 EHRs including Epic, athenahealth, eClinicalWorks, and ModMed. There is no rip-and-replace, it goes live in days, and every resolved patient interaction is what we charge for.\r\n\r\nOur starting point is healthcare. The mission is to rebuild the critical industries America runs on, including government, finance, insurance, and manufacturing, by deploying AI agents into the slow, manual workflows at their core.", "one_liner": "Healthcare AI", "team_size": 2, "industry": "Healthcare", "subindustry": "Healthcare -> Healthcare IT", "launched_at": 1785365250, "tags": [ "Healthcare", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare", "Healthcare IT" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/elite", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/elite.json", "year_founded": 2026, "city": "New York City", "country": "US", "linkedin_url": "https://www.linkedin.com/company/this-is-elite", "x_url": "https://x.com/Stratum_Inc", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/950a2e88427ff9cd0ec85f5da8c2e8bd7d8333df.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/c94ced9719faa591972f4a2547e746ac79c217b8.png", "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/ada3710bf9074942e19405ec614b7911439b6ca1.png" ], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1648613, "full_name": "Raj Jagiasi", "title": "Founder/CEO", "founder_bio": "Co-founder @ Elite - Rebuilding critical American industries. Prev at Deel. College dropout.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/raj-jagiasi/", "x_url": "https://x.com/raj_jagiasi", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/136a9babd34f78ebf558618f247fba37378307d3.jpg" }, { "user_id": 1303141, "full_name": "Vihaan Jagiasi", "title": "Founder/CTO", "founder_bio": "co-founder @ Elite. engineer #3 at cadastral (acq by Legora). waterloo dropout", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/vihaanjagiasi/", "x_url": "https://x.com/VihaanJagiasi", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/32665114b168ae0427fec8e263c5cb49bd0d84c4.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108362, "title": "Stratum Industries - Applied AI for Governments", "tagline": "AI agents that automate the complex manual workflows governments run on.", "body": "**TL;DR: Stratum builds AI agents that automate the slow, manual processes governments run on. We're starting with application review - clearing backlogs in a fraction of the time and saving agencies tens of billions a year.**\n\n![uploaded image](/media/?type=post&id=108362&key=user_uploads/1648613/f70bb601-955f-4c6c-946f-8d8c90d04fa6)\n\n**The problem**\n\nGovernments run on manual work. Millions of complex processes - reviews, approvals, filings, claims -are still handled by hand, one file at a time, on software built long before AI existed.\n\nTake application review. Every permit, license, patent, drug approval, and benefits claim in America is reviewed by hand, one file at a time. Millions sit in queues while people wait months and years: veterans on disability ratings, families on pensions, builders on permits.\n\nThe tools reviewers open every morning were built long before AI. So the work stays manual, and the queue only grows.\n\nThis isn't one broken workflow. It's how the entire government runs.\n\n**What Stratum does**\n\nStratum builds AI agents that automate the slow, complex, manual processes governments run on.\n\nWe're starting with application review.\n\nA reviewer uploads an application and selects the rubric to check it against. From there, an agent takes over: it plans the review, works through the file end to end, and checks every requirement, routing each step to the right model.\n\nA human approves every outcome, and every step is fully auditable.\n\nThe difference is scale. Instead of one reviewer on one file, hundreds of agents run in parallel — clearing an entire backlog in the time it once took to review a single application.\n\n![uploaded image](/media/?type=post&id=108362&key=user_uploads/1648613/e23f9409-271f-4879-95aa-374eb37ae64f)\n\n**Who we are**\n\nWe're Vihaan (VJ) and Raj - brothers.\n\nVihaan (CTO): founding engineer #3 at Cadastral (acquired by Legora), waterloo dropout.\n\nRaj (CEO): led GTM at Deel, dropped out of CS grad school at the University of Bath.\n\n**Where this goes**\n\nApplication review is the first workflow. The mission is to integrate AI across every vertical inside the government.\n\nGovernments are among the largest, most complex organizations on earth — and almost none of it has been touched by modern AI. We're changing that, saving agencies tens, eventually hundreds, of billions a year and making the government one of the most efficient institutions in the world.\n\nLooking to bring AI into your agency? Reach out to the team or book a demo [stratumindustries.co](https://stratumindustries.co/)\n\nContact: [founders@stratumindustries.co](mailto:founders@stratumindustries.co)", "url": "https://www.ycombinator.com/launches/SBm-stratum-industries-applied-ai-for-governments", "votes": 9, "created_at": "2026-08-14T18:00:00.843Z" }, { "id": 108348, "title": "Stratum Industries - clearing the entire U.S. Government backlog", "tagline": "AI agents that review government applications end-to-end, so backlogs that took years clear in minutes.", "body": "**TL;DR**\n\nStratum builds AI agents that clear the U.S. government's backlog. A reviewer drops in an application, picks the rubric to check it against, and an agent plans, executes, and reviews the file end to end. Hundreds run in parallel with a human approving every decision, so a backlog that used to take months clears in the time it took to review one file.\n\n**The Problem**\n\nEvery permit, license, patent, drug approval, grants and benefits claim in America has to be reviewed by a government adjudicator, on software built decades before AI. The result is millions of applications stuck in queues, and real people waiting months or years: veterans on disability ratings, families on pensions, builders on permits. The approval isn't the problem. The software it runs on is.\n\n**How Stratum Works**\n\n1. A reviewer drops in an application and selects the rubric it needs to be checked against.\n2. An agent reads every page and checks the file against that rubric.\n3. Every step routes to the model that reviews it most accurately for the least cost.\n4. Anything missing goes back to the applicant as one specific request.\n5. The officer gets the whole file summarized, with every citation behind it, and approves or denies it.\n\nHundreds of these run in parallel in the cloud, with full auditability at every step, and a person always makes the final call.\n\n![uploaded image](/media/?type=post&id=108348&key=user_uploads/1648613/15e81a7e-32ca-4280-9822-bc49caed3028)\n\n**The Vision**\n\nIn a couple of years, dealing with the government should feel as frictionless as booking an Uber or paying with Apple Pay. You tell your Stratum agent what you need, a permit, a license, a claim, and it fills out the application and submits it. On the other side, the government's Stratum agents review it, ask your agent for anything missing, and settle it between themselves. The decision comes back the same day, in minutes. No waiting weeks or months. Clearing today's backlog is only the beginning.\n\n**Team**\n\nWe're Vihaan and Raj, brothers. Vihaan was founding engineer #3 at Cadastral (acquired by Legora) and dropped out of Waterloo to build Stratum. Raj led GTM at Deel, averaging 220% of quota over two years and ranking top 5 of 1,000+ reps globally, and dropped out of grad school at Bath to build Stratum.\n\n![uploaded image](/media/?type=post&id=108348&key=user_uploads/1648613/40ac0440-bf77-46cf-b30e-370f734c3203)\n\n**Our Ask**\n\n* Do you work in government and struggle with a review backlog? We want to talk. Email [founders@stratumindustries.co](mailto:founders@stratumindustries.co)\n* Know someone who runs adjudication at a federal, state, or local agency? Intro us.", "url": "https://www.ycombinator.com/launches/SBY-stratum-industries-clearing-the-entire-u-s-government-backlog", "votes": 2, "created_at": "2026-07-30T01:21:28.889Z" } ] }, { "id": 32525, "name": "FlowManual", "slug": "flowmanual", "former_names": [ "ProManual" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/f7ea500c797163e99dad11a205953572d46a0034.png", "website": "https://flowmanual.com/", "all_locations": "San Francisco, CA, USA", "long_description": "FlowManual is the all-in-one AI estimating, bidding, and purchasing platform for both general and MEP construction contractors.\r\n\r\nBuilt by two engineers and Harvard dropouts, FlowManual automates the buying process during pre-construction. With FlowManual, specialty contractors win more work while general contractors deliver at higher standards.", "one_liner": "AI for Construction's Back Office", "team_size": 2, "industry": "Real Estate and Construction", "subindustry": "Real Estate and Construction -> Construction", "launched_at": 1782366913, "tags": [ "SaaS", "Construction", "Workflow Automation" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Real Estate and Construction", "Construction" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/flowmanual", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/flowmanual.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/flowmanual/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/f7ea500c797163e99dad11a205953572d46a0034.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 3655483, "full_name": "David Shijoon Bae", "title": "Founder/CEO", "founder_bio": "Co-founder & CEO @ FlowManual. Automating Construction Contract Procurement. Previously MechE @ Harvard.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shijoonbae/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d1f8cb67b8dff98305776484a9dd815171923863.jpg" }, { "user_id": 3054292, "full_name": "Michael Lin", "title": "Founder/CTO", "founder_bio": "Co-Founder and CTO at FlowManual. Previously SWE intern at Dijie, where he built a video-based landing page that scaled revenue from $3K to $336K MRR and reduced latency from 800ms to 100ms. Published CS researcher at UC Santa Cruz, co-authored paper submitted to IEEE IV 2025.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/itsmichaellin", "x_url": "https://x.com/immichaellin", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/7fbb62b7b30790e8c248f3f80ac177a3bb3c451c.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109738, "title": "FlowManual — An AI-native back office for construction contractors", "tagline": "AI bidding, estimating, and purchasing for construction contractors, clearing risk without blind AI reliance.", "body": "Hey everyone! We’re David and Michael, co-founders of FlowManual.\n\n**What FlowManual is**: AI preconstruction for General and MEP contractors, executing projects while cutting risk. \n\n**TL;DR:** Construction contractors waste hundreds of hours monthly on manual tasks, sifting through 500-page PDFs only to miss fine print that can cost tens of thousands. FlowManual automates the back-office work before and during construction, so teams can reliably secure and complete projects faster than ever before.\n\n\\\n\\\n**The Problem:**\n\nMeet Jeremy. He’s a project manager at a mid-sized contractor, and his job is to secure projects and then make sure the company doesn’t lose money on already thin margins.\n\nA bid invite lands in his inbox. To put together one number his firm can send in, Jeremy has to read thousands of pages. A 400-page project manual, drawings, insurance, spec sections, and so much more. Buried somewhere inside all this print is a line that can ultimately decide whether this job makes or loses money. From a liquidated-damages clause to a scope item sneakily handed to the mechanical contractor, if Jeremy misses it, it’s a $5K-$10K mistake.\n\nJeremy’s got one week, three other bids on his desk, and no way to read all of it as carefully as he’d like. So he reads what he can and sends in a number.\n\nSay he wins. A change in the signed contract by the General Contractor or a sneaky difference in vendor pricing costs the company thousands after the fact. Every step of the way requires hundreds of pages of reading and referencing done under a deadline. And Jeremy does this for every job.\n\n**The Solution:**\\\nFlowManual automates these historical workflows that contractors run on, allowing teams to submit bids and complete projects faster than ever before. Here’s a couple major features:\n\n* AI Takeoffs from drawings\n* Drawing comparisons / overlay\n* Bid leveling\n* Contract Review\n\nOur software doesn’t do the job for you, but it works alongside you to reduce time spent and mistakes made.\n\n**About Us:**\\\nMichael and David were college roommates and engineers at Harvard. When renovations for a heater in our dorm took 6 weeks, we poked at the problem, realizing it stems from a larger issue. Through hundreds of conversations with professionals, we found an industry that ran on file cabinets, Excel, and page-by-page manual review.\\\n\\\n**Our Ask:**\n\n* **Know anyone in construction?** Connections to anyone in the industry, including on the legal and tech side, is greatly appreciated. Anyone in Project Management, Business Development, Marketing, IT are ideal! Reach us at [founders@flowmanual.com](mailto:founders@flowmanual.com) \n* **Feedback, intros, or just want to chat?** Book a call with us, or email [founders@flowmanual.com](mailto:founders@flowmanual.com)!", "url": "https://www.ycombinator.com/launches/SXy-flowmanual-an-ai-native-back-office-for-construction-contractors", "votes": 5, "created_at": "2026-08-10T13:00:00.161Z" } ] }, { "id": 32526, "name": "Pluto", "slug": "talentpluto", "former_names": [ "Talentpluto" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/5ed12c5974a12759a4a83154066015d6c91fc4df.png", "website": "https://talentpluto.com", "all_locations": "San Francisco, CA, USA", "long_description": "Resumes made careers readable. LinkedIn made them filterable. Pluto makes them searchable.\r\n\r\nProfessionals spend 10 minutes talking to Pluto's voice agent about what they've done, what they're great at, and what they want next. Pluto turns that into a structured picture of who someone actually is, capturing the competency and intent that resumes and LinkedIn don't.\r\n\r\nWhen companies and AI agents go looking to hire, Pluto makes the right people discoverable.\r\n\r\n27K+ professionals have shared their career context with Pluto, and 50+ companies, including Mercor, Warp, and Rho, use it to find and reach the right people. Long-term, we're building the LinkedIn for agents.\r\n", "one_liner": "LinkedIn for agents", "team_size": 5, "industry": "B2B", "subindustry": "B2B -> Recruiting and Talent", "launched_at": 1782230867, "tags": [ "AI", "Careers", "Conversational AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Recruiting and Talent" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/talentpluto", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/talentpluto.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/talentpluto-inc", "x_url": "https://x.com/talentpluto", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/talentpluto", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/5ed12c5974a12759a4a83154066015d6c91fc4df.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 1666957, "full_name": "Sahil Seth", "title": "Founder/CEO", "founder_bio": "CEO @ Pluto. Previously built an AI interviewer used by billion-dollar recruiting firms. Led AI recruiting transformations at Bain and advised a Fortune 5 company. Graduated summa cum laude from Northwestern, elected to Phi Beta Kappa, and received the Frederick Shipp Diebler Award as the top economics undergraduate.", "is_active": true, "linkedin_url": "https://linkedin.com/in/sahilseth9", "x_url": "https://x.com/sahil_seth9", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b1ef8412e27d4d136715fadcd78e698768fe50aa.jpg" }, { "user_id": 3207785, "full_name": "Maki Potamianos", "title": "Founder/CPO", "founder_bio": "CPO @ Pluto. Founded and scaled an edtech marketplace in college. Sole graduate with a dual degree in both applied mathematics and finance, completing the mathematics degree in two years.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/michaelpotamianos/", "x_url": "https://x.com/Maki_Potamianos", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3bd6ff8b549fd4ce443e3677fd9844e3a2a11cdf.jpg" }, { "user_id": 1155860, "full_name": "Tony Xavier", "title": "Founder/CTO", "founder_bio": "CTO @ Pluto. Built an open source no-code editor competing with Lovable and Replit; researched voice agents and RAG systems. Placed top 0.1% nationally in India in math and physics. UMich Physics, Mechanical Engineering, and CS.\n", "is_active": true, "linkedin_url": "https://linkedin.com/in/tony-xavier", "x_url": "https://x.com/TonyXavier_", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/51b490006f7200aec10a5f9aafd7cdab562462ea.jpg" } ], "jobs": [ { "id": 115910, "title": "Founding Engineer", "url": "https://www.ycombinator.com/companies/talentpluto/jobs/SIrtJNc-founding-engineer", "location": "San Francisco, CA, US / New York, NY, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$120K - $200K", "equity_range": "0.15% - 1.00%", "experience": "1+ years", "visa": "US citizen/visa only", "skills": [ "JavaScript", "PostgreSQL", "React", "TypeScript", "SQL", "Next.js", "LLMs" ] } ], "news": [], "launches": [ { "id": 105611, "title": "Pluto - AI voice agent that helps professionals get discovered", "tagline": "Pluto is an AI voice agent that learns a professional’s story in a 10-minute conversation, makes them discoverable to the right people and AI agents, and creates warm introductions to startups when there’s a fit.", "body": "Hey YC, we’re Sahil, Maki, and Tony. We’re building Pluto ([talentpluto.com](https://talentpluto.com/))\n\nResumes made careers readable. LinkedIn made them searchable. Pluto makes them understood.\n\n\n\n**TL;DR** - A professional talks to Pluto for 10 minutes about what they’ve done, what they’re great at, and what they want next. Pluto makes that context discoverable to the right people and AI agents, helping them understand the person more fully. When there’s a real fit, Pluto can also make a warm introduction to a startup.\n\n18K+ professionals have already shared their career context with Pluto. 100+ companies use Pluto to meet people with their full story, including Mercor, David AI, Rho, Warp, and Runlayer.\n\n**Ask:** Talk to Pluto and let it get to know your work. If you’re hiring, send us the role and we’ll introduce you to people with context and confirmed interest.\n\n### **The problem**\n\nA resume and online profile are not enough for the AI era.\n\nThey show where someone worked and what title they had. They do not capture what the person actually did, what they are unusually good at, and what they want next.\n\nThat context usually lives in conversations. So professionals get reduced to keywords, logos, and titles, while the people and AI agents trying to understand them are left with an incomplete picture. \n\nHiring makes this problem obvious. Companies miss great people because the real signal never makes it online. But as AI agents increasingly help people find, evaluate, and connect with one another, every professional will need a richer, more accurate way to be understood.\n\n### **The solution**\n\nPluto is an AI voice agent that understands a professional’s career and puts that understanding to work.\n\nProfessionals talk to Pluto for 10 minutes about their experience, strengths, goals, preferences, and what they want to keep private. Pluto asks follow-up questions, remembers the context missing from their resume, and makes them discoverable to the right people and AI agents.\n\nCompanies tell Pluto who they want to meet. Pluto searches its network and the broader market, understands fit on both sides, privately confirms interest, and makes a warm introduction when both sides want to connect. No mass applications. Professionals stay private and in control. Companies meet people who are already understood.\n\n**How It Works**\n\n1. Talk to Pluto for 10 minutes\n2. Review what it learned and choose what to share\n3. Get found and understood by the right people and AI agents\n4. Get introduced to startups when there’s a fit\n\n### **Why We Started This**\n\nThe best hiring has always happened through trusted, high-context introductions. We think AI can make that possible for everyone.\n\nWarm introductions are one way Pluto puts that understanding to work today. Long term, Pluto becomes the career context layer for the AI era - the version of a professional’s story that agents can understand, trust, and act on, while the professional stays in control.\n\n### **Our Ask**\n\nIf you’re a professional, talk to Pluto and let it get to know your work. \n\nIf you’re a founder, hiring leader, or talent engineer, send us the role you’re hiring for.\n\n[talentpluto.com](http://talentpluto.com)\n\n\\- Sahil, Maki, Tony", "url": "https://www.ycombinator.com/launches/RTP-pluto-ai-voice-agent-that-helps-professionals-get-discovered", "votes": 19, "created_at": "2026-07-16T16:33:18.764Z" } ] }, { "id": 32527, "name": "machine0", "slug": "machine0", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/0eb24a85015d2fd30115c143876496fed0a17210.png", "website": "https://machine0.io", "all_locations": "San Francisco, CA, USA", "long_description": "machine0 is a CLI for agent compute: `machine0 new mybox` gives your agent a full cloud VM, billed by the minute, from small up to 60 vCPU / 240 GB and 8×H100s. Agents drive it themselves via CLI or MCP.\r\n\r\nAgent workloads are moving from ephemeral to always-on. A coding agent working on a hard feature runs 6-8 hours. Auto-research agents run for days. machine0 is the environment in which these agents run.", "one_liner": "AWS for Agents", "team_size": 1, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1780987913, "tags": [ "Developer Tools", "Cloud Computing" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/machine0", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/machine0.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/machine0", "x_url": "https://x.com/machine__0", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/0eb24a85015d2fd30115c143876496fed0a17210.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 31731, "full_name": "Barnaby Malet", "title": "Founder/CEO", "founder_bio": "Engineer. 2 x YC Founder. Currently founder/CEO of machine0 (S26). Previously cofounder/CTO/CPO of Upflow (W20). Before all that, Computer Science @ Imperial College London.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/barnabymalet/", "x_url": "https://x.com/barnabymalet", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3aa5cc9ddb2b21b5e10a21bd90ac20c4978f083a.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108327, "title": "machine0: Cloud computers for AI agents", "tagline": "One command spins up a persistent VM. Your agent drives it via CLI or MCP, building its own fleet from 1 vCPU up to 60 vCPU / 240 GB and 8×H100s.", "body": "TLDR: machine0 is a CLI for agent compute: “machine0 new mybox” gives your agent a full cloud VM, billed by the minute, from $0.013/hr up to 60 vCPU / 240 GB and 8×H100s. Agents drive it themselves via CLI or MCP.\n\n\n\n## The Pain\n\nAgent workloads are moving from ephemeral to always-on. A coding agent working on a hard feature runs 6-8 hours. Auto-research agents run for days. OpenClaw & Hermes run 24/7. As you run more in parallel:\n\n* **Resources**: a few agents on a big codebase saturate your RAM and CPU. Auto-research needs GPUs you don't have.\n* **Security**: “—yolo” on your personal machine is one prompt injection away from exfiltrated credentials.\n* **Availability**: close your laptop and the agent dies mid-task.\n* **Isolation**: there's no clean line between everything on your laptop and the minimum this agent actually needs.\n\n## The Solution\n\nmachine0 gives every agent its own computer. It's a CLI that both humans and agents can use without reading docs:\n\n* “machine0 new mybox” creates an SSH-ready VM with a static IP and HTTPS endpoint. Always on (99.99% uptime) until you switch it off.\n* **Billed by the minute.** 1 vCPU / 1 GB at $0.013/hr up to 60 vCPU / 240 GB, plus GPUs from RTX 4000 Ada to 8×H100s.\n* **Profiles.** Bundles of MCP connections, credentials, prompts, and env vars, injected at VM creation. Each agent gets exactly the capabilities (or personality) you choose, and nothing else.\n* **Agents self-serve.** Hand the CLI or MCP server to Claude, Codex, or OpenCode and it manages its own fleet: spin up a box for a build, snapshot it, tear it down.\n* **Suspend & resume.** Pause an idle machine, stop paying for compute, resume exactly where it left off.\n* **Snapshot & clone.** Configure one machine perfectly, then stamp out ten identical agents from it.\n* **NixOS is first-class.** Declarative, reproducible machines.\n\n## How people use it\n\n* **Compute for your agent**: give it the CLI and it spins up its own VMs for builds, tests, and long jobs instead of grinding your laptop.\n* **Hosting the agent itself**: run Claude, Codex or Hermes 24/7. Snapshot once, then clone as many as you want.\n* **Swarms of agents**: an orchestrator agent spawns worker machines, farms out subtasks, and tears them down when the work is done.\n* **Auto-Research on GPUs**: drop an agent on an H100 with a goal and let it run for days.\n\n## About me\n\nI'm a solo, second-time YC founder. I was CTO of Upflow (YC W20), where I scaled EPD post series A and personally owned DevOps, infra, and security for years because they were too high-stakes to delegate. Upflow handles sensitive accounting data for companies like Postman, ProductBoard, and Lattice, and is doing many millions in ARR.\n\nBefore that I built consumer payments at Hike Messenger (100M+ users) in India. Working with sensitive data and infra for years is what got me deep into the problems machine0 solves, security especially.\n\n## Asks 🙏\n\n* Try it: \n* Let's talk about your agent fleets, software factory or auto-research: [hello@machine0.io](mailto:hello@machine0.io)", "url": "https://www.ycombinator.com/launches/SBD-machine0-cloud-computers-for-ai-agents", "votes": 22, "created_at": "2026-07-31T15:30:36.400Z" } ] }, { "id": 32551, "name": "Bloomy", "slug": "bloomy", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/209c1fdd9d657cb4fad05300bbb0faf304f4d29c.png", "website": "https://bloomylearning.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Bloomy is the best way to teach anything in school. We provide schools with an AI-native academic system comprising core and supplemental curriculum, AI teaching, and assessment to help K-12 students learn faster. Backed by the founders of Clever.", "one_liner": "AI-native learning for every school", "team_size": 1, "industry": "Education", "subindustry": "Education", "launched_at": 1782182610, "tags": [ "AI-Enhanced Learning", "Education", "eLearning", "Personalization", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Education" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/bloomy", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/bloomy.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/bloomy-yc-s26/", "x_url": "https://www.x.com/bloomylearning", "fb_url": "https://www.facebook.com/profile.php?id=61590258284564", "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/209c1fdd9d657cb4fad05300bbb0faf304f4d29c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 3410278, "full_name": "Alex Southmayd", "title": "Founder/CEO", "founder_bio": "Founder of Bloomy. Previously 7th grade teacher with Teach For America; GMAT/GRE instructor at Manhattan Prep; Head of Driver Acquisition for Lyft New England (built an acquisition playbook that scaled nationally); Engagement Manager at McKinsey & Company (strategy, VC/PE diligence, sales acceleration). BA from Amherst College, M.Ed. from BU, MBA from Stanford GSB", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/alex-southmayd/", "x_url": "https://x.com/alexsouthmayd", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b72aa9651889f097ba934a5cdd6fab522ce33aa7.jpg" } ], "jobs": [ { "id": 100730, "title": "Founding Engineer", "url": "https://www.ycombinator.com/companies/bloomy/jobs/7Kz07Zt-founding-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$175K - $200K", "equity_range": "1.50% - 2.00%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [] }, { "id": 116273, "title": "Chief Technology Officer", "url": "https://www.ycombinator.com/companies/bloomy/jobs/P5qe6yZ-chief-technology-officer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$200K - $250K", "equity_range": "2.00% - 3.00%", "experience": "6+ years", "visa": "US citizen/visa only", "skills": [] }, { "id": 116274, "title": "Partnerships Account Executive", "url": "https://www.ycombinator.com/companies/bloomy/jobs/e8MhnxJ-partnerships-account-executive", "location": "US / Remote (US)", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "$200K - $210K", "equity_range": "", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [] }, { "id": 116275, "title": "Partnerships Account Manager", "url": "https://www.ycombinator.com/companies/bloomy/jobs/FQRMUCS-partnerships-account-manager", "location": "US / Remote (US)", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "$200K - $210K", "equity_range": "", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 104376, "title": "Bloomy: AI-powered mastery learning for K-12", "tagline": "We use AI to help kids learn faster", "body": "**Hi everyone - I'm Alex, founder of Bloomy.**\n\n**TL;DR: We help kids learn faster by using AI. In early pilots, students grew \\~2x faster than the national average.**\n\n[Bloomy](https://bloomylearning.com) is an AI-powered mastery learning platform for K-12 students. We diagnose students’ skill gaps, place students on a personalized learning path, and give them standards-aligned lessons and a Socratic AI tutor that helps without just giving away the answer.\n\nLaunch video: \n\n**Why I'm building this**\n\nI started as a teacher. I taught 7th grade English and Writing with Teach For America, and every day I struggled to deliver differentiated instruction to 30 students with 30 different sets of needs.\n\nSome students needed remediation, some needed acceleration, and many needed a tutor sitting next to them helping them reason through the next step. Benjamin Bloom's 2-sigma result always felt obviously true to me; the hard part was making that kind of tutoring affordable and available to every child.\n\nThen AI changed the cost curve. \n\nWhen I saw innovative school concepts such as Alpha School, the model clicked: mastery-based learning, personalization, and high-agency students. But I kept thinking about the families and schools that already exist: homeschool families, microschools, hybrid schools, and regular classrooms where most kids are today.\n\n**What Bloomy does**\n\nBloomy starts with a diagnostic (we integrate with third-party diagnostics and provide our own in-house one) and places each student onto a personalized learning path. Students work one skill at a time, get a short, standards-aligned lesson, practice, ask BloomyBot for Socratic help when they're stuck, and only move forward after showing 90%+ mastery. The skill path updates live based on the student’s performance and our own knowledge graph.\n\n![uploaded image](/media/?type=post&id=104376&key=user_uploads/3410278/aff91335-737d-4b76-bcc2-40414d3f3c08)\n\n_Our knowledge graph informs skill progression to make learning more efficient._\n\nParents and teachers can see what the student has mastered, what's in progress, and where they need support.\n\n**Where we are**\n\nWe're being adopted in many different contexts: districts with traditional schools, charter schools, hybrid schools, microschools, homeschools, and even families who just want some extra academic support for their kids.\n\nFamilies can subscribe to Bloomy monthly or annually, and families using ESA/EFA funds may also qualify for direct pay or reimbursement depending on their state’s rules.\n\nBefore Bloomy, I taught with TFA, earned my Master’s in Education, designed GMAT and GRE curriculum for Manhattan Prep, led Lyft’s driver growth team for New England, earned an MBA from Stanford, and managed AI transformation engagements at McKinsey. Bloomy is the thing that combines all the aspects of my career: learning design, go-to-market, and AI.\n\nIf you want to reach out, you can email me at [alex@bloomylearning.com](mailto:alex@bloomylearning.com) or follow our updates on [LinkedIn](https://www.linkedin.com/company/bloomy-yc-s26/).\n\n![uploaded image](/media/?type=post&id=104376&key=user_uploads/3410278/9c55515e-0577-460d-955d-e1c3cb587f06)\n\n", "url": "https://www.ycombinator.com/launches/R9U-bloomy-ai-powered-mastery-learning-for-k-12", "votes": 10, "created_at": "2026-07-01T15:05:35.719Z" } ] }, { "id": 32552, "name": "6thSense", "slug": "6thsense", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/24cb943ce2196a35afe2b7d03f61eab2d4ec937a.png", "website": "https://6thsense.dev", "all_locations": "San Francisco, CA, USA", "long_description": "", "one_liner": "Nervous System for Physical AI", "team_size": 4, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1782449364, "tags": [ "Robotics" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/6thsense", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/6thsense.json", "year_founded": null, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/6thsenseai", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/24cb943ce2196a35afe2b7d03f61eab2d4ec937a.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/e4fdc2d7bcabc67030a4042e4ff79a3d9e187aea.png" ], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 3429291, "full_name": "James Baek", "title": "Founder/CEO", "founder_bio": "Founder & CEO at 6thSense. Founding member at Ibebu (Series C). Security Specialist, US Army (USFK). Exited a six figure content creation business and built numerous faceless brands. BME & Biochem at Georgia Tech. Growth Hacker\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/james-baek1/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1cb03c9cf75088d5d10002833f9baf36b05998c2.jpg" }, { "user_id": 1407900, "full_name": "Ronak Agarwal", "title": "Founder/CTO", "founder_bio": "Founder & CTO at 6thSense. ML at Amazon. Production infrastructure at DoorDash. Built autonomous drone delivery at Chick-fil-A. Built several startups before 6thSense, including Qard. Georgia Tech CS & Economics graduate (Intelligence + Modeling & Simulation); left a CS master's in Data Science & Systems Architecture. AI/ML & Systems Builder.", "is_active": true, "linkedin_url": "https://linkedin.com/in/ronak-agarwal9", "x_url": "https://x.com/ronakgt", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/338442a428cd26e2dd42028368b6924295dff73a.jpg" }, { "user_id": 3367434, "full_name": "Alex Hyungwoo Noh", "title": "Founder/COO", "founder_bio": "Founder at 6thSense. Bio/ML scholarship from Samsung at age of 17. UChicago CS/Econ. Ex-Army Recon. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/hyungwoonoh/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f41a76394905402b8f244c67a750e1792d0e33dd.jpg" }, { "user_id": 2055952, "full_name": "Matt Wulff", "title": "Founder/CPO", "founder_bio": "Co-Founder of 6thSense (YC S26). Lead Vision Guided Robotics (ME) @Tesla at 20yo, Homeschooled, First Company at 16yo, Build Rockets for fun, Lead Tactile Data Capture at Mecka AI, Dreamer and adrenaline junky", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/makeitmatto/", "x_url": "https://x.com/makeitmatto", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1fce472f4227cba7507f4caf1b2de54916ed016b.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 104007, "title": "6thSense — Giving Robots Touch", "tagline": "Giving Robots Touch", "body": "6thSense builds wearable tactile sensing rigs that capture how human hands interact with the world — pressure, contact, grip, and pose — across touch, vision, depth, force, and 3D hand pose, all in real environments. We sell that multimodal data to the companies training the next generation of dexterous robots.\\\n\\\n\n\nRobots are getting good at seeing. They're still terrible at touching. That's because there's almost no data capturing what a hand actually feels while doing real tasks in real homes. Lab data doesn't cut it. We go anywhere and capture the missing layer that manipulation models need to learn fine motor control.\n\nWhat makes our data different is the hardware we built to collect it. Our tactile sensing devices are designed and manufactured in-house, purpose-built to capture the richest touch signal in the industry — with scale in mind from the very beginning, enabling us to actually solve the tactile data problem. Every sensor maps directly to how robotic skin needs to work: a 1-to-1 analog for the tactile arrays going into the next generation of robot hands. That means the data we collect today doesn't just train better policies. It transfers directly to the hardware it's meant to run on. The team comes from Tesla, Mecka, Samsung, DoorDash, and Amazon, with expertise across hardware, sensors, ML, robotics, and software at scale. If you're training a manipulation policy or need custom tactile data, reach out.", "url": "https://www.ycombinator.com/launches/R3X-6thsense-giving-robots-touch", "votes": 7, "created_at": "2026-07-05T16:58:56.478Z" } ] }, { "id": 32553, "name": "Nori", "slug": "noril1", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/4b7cc06bfd9d08d6eae3bc20eca8c056f893bf8b.png", "website": "http://norirobotics.com", "all_locations": "San Francisco, CA, USA", "long_description": "We manufacture Nori: A sub-$2000 bimanual mobile manipulator that can clean rooms, restock shelves, and fold clothes. Nori can handle basic tasks out of the box, and its software lets anyone teach Nori to automate work specific to their home or business. \r\n\r\nNori aims to solve the robotics data bottleneck. By deploying affordable robots to customers around the world, we can gather the diverse data required to train the next generation of generalist robotic policies. Every user of Nori makes Nori smarter.\r\n\r\nSince launching 6 weeks ago we have already deployed our first robot and have over $440k+ in sales.\r\n\r\nWe can bring general-purpose robotics to homes and businesses that could never afford it before. There are 132.7 million households and 8.4 million employer business locations in the U.S. At $1,688 per robot, selling just one Nori to 0.5% of them represents a $1.2B market opportunity in the U.S. alone. ", "one_liner": "The most capable American-made humanoid robot for $1,688", "team_size": 5, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1783099451, "tags": [ "Hard Tech", "Hardware", "Robotics" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": [ { "question": "What is the core problem you are solving? Why is this a big problem? What made you decide to work on it?", "answer": "Robots are finally capable enough to be useful, but they're priced like cars, not appliances. A capable humanoid costs $20,000+, so general-purpose robots stay in labs and pilots instead of homes. \r\n\r\nIt's a big problem because household labor is one of the largest unpriced markets in the world. Today's answer is a stack of single-task machines, a dishwasher, a washer, a vacuum, each $600 to $1,000 and doing exactly one thing. Whoever ships a general-purpose robot at appliance prices doesn't win a niche, they open consumer robotics the way the PC opened computing.\r\n\r\nWe're robotics researchers from Columbia (RoboPIL, CGUI). After years watching \\[$100K research robots struggle with tasks cheap hardware could handle\\], we realized the gap between lab robots and home robots isn't intelligence, it's cost discipline. Nobody was going to fix that from inside a lab, so we left to build it: a $1,688 robot anyone can teach." } ], "url": "https://www.ycombinator.com/companies/noril1", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/noril1.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/norirobotics/", "x_url": "https://x.com/NoriRobotics", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/4b7cc06bfd9d08d6eae3bc20eca8c056f893bf8b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 2661900, "full_name": "Antonio Sitong Li", "title": "Founder/CTO", "founder_bio": "Antonio is the founder of Nori Robotics. He studied Computer Science and Architecture at Columbia, where he held research fellowships with the Laidlaw Scholars Program, the Data Science Institute, and the Graphics & UI Lab. His research focused on teaching robots new tasks through VR demonstrations.\n\nHe also holds a national patent for a computer-vision system, and previously co-founded Truely, which grew to 2,000+ users and 1.5M+ impressions.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/antonio-sitong-li/", "x_url": "https://x.com/AntonioSitongLi", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/31394af155bdf33ce5bb832507516ea31b4407bd.jpg" } ], "jobs": [ { "id": 117560, "title": "Data Operations Lead", "url": "https://www.ycombinator.com/companies/noril1/jobs/njoB5k2-data-operations-lead", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "$100K - $150K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 117561, "title": "Full-Stack Engineer", "url": "https://www.ycombinator.com/companies/noril1/jobs/cERiSKM-full-stack-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$100K - $150K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [] }, { "id": 117562, "title": "Multimodal Engineer", "url": "https://www.ycombinator.com/companies/noril1/jobs/boIEFcs-multimodal-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Machine learning", "salary_range": "$130K - $200K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [] }, { "id": 117563, "title": "VLM Researcher", "url": "https://www.ycombinator.com/companies/noril1/jobs/3VxHqDs-vlm-researcher", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Machine learning", "salary_range": "$150K - $250K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [] }, { "id": 117564, "title": "LLM Researcher", "url": "https://www.ycombinator.com/companies/noril1/jobs/2LGcxo3-llm-researcher", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Machine learning", "salary_range": "$150K - $250K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [] }, { "id": 117565, "title": "Applied AI Engineer, Agents & Robot Learning", "url": "https://www.ycombinator.com/companies/noril1/jobs/ZSfyL0x-applied-ai-engineer-agents-robot-learning", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Robotics", "salary_range": "$150K - $200K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [] }, { "id": 117566, "title": "Design Engineer", "url": "https://www.ycombinator.com/companies/noril1/jobs/XrKBJq1-design-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Design", "role_type": "Product design", "salary_range": "$100K - $130K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [] }, { "id": 117567, "title": "Social Media & Content Marketer", "url": "https://www.ycombinator.com/companies/noril1/jobs/jVWzI5j-social-media-content-marketer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Marketing", "role_type": null, "salary_range": "$35 / hourly", "equity_range": "", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 117568, "title": "Electrical Engineer", "url": "https://www.ycombinator.com/companies/noril1/jobs/Exs6QKi-electrical-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Electrical", "salary_range": "$110K - $120K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [] }, { "id": 117569, "title": "Robot Assembly & Automation Technician", "url": "https://www.ycombinator.com/companies/noril1/jobs/qciFkKu-robot-assembly-automation-technician", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "$47.8K - $58.2K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] } ], "news": [ { "title": "American Humanoid Robot Launches For Just $1,688, Delivery This Fall, Made In San Francisco", "url": "https://www.forbes.com/sites/johnkoetsier/2026/08/18/american-humanoid-robot-launches-for-just-1688-delivery-this-fall-made-in-san-francisco/", "date": "Aug 20, 2026" } ], "launches": [ { "id": 111315, "title": "Nori: The $1688 humanoid robot", "tagline": "Most humanoids cost 16k, ours cost 1.6. 19 DOF, 1.5 KG lift per arm", "body": "Hey YC! 👋\\\n\\\nI’m Antonio from Nori Robotics\n\nWe manufacture humanoids in SF, and we do it at 10% the cost of the competition \n\n**Our ask:** If you are doing any development related to robotics and you think that Nori can be useful, fill out this [form](https://docs.google.com/forms/d/e/1FAIpQLSdcHJxwXCBo5dz7IbIQ30akCu5JaD5Tbv5dP9gIQxvbJQR-3Q/viewform?usp=dialog) or reach out to me on: [X](https://x.com/AntonioSitongLi) or [linkedin](https://www.linkedin.com/in/antonio-sitong-li/). \n\nHere is a video demo of what Nori can do 👇\n\n\\\n\\\nLearn more at [norirobotics.com](http://norirobotics.com)\n\n\\\n**Nori Specs:**\n\n19 DOF\n\n\n\n55 kg linear lift\n\n\n\n1.5 kg payload per arm (2 arms)\\\n\n\nMicrophone array + Speaker (full-duplex speech dialogue)\n\nRpi 5 4GB compute module \n\n4 \\* 720p 30 fps camera", "url": "https://www.ycombinator.com/launches/SxP-nori-the-1688-humanoid-robot", "votes": 16, "created_at": "2026-08-20T18:30:00.323Z" } ] }, { "id": 32554, "name": "Ooak Data", "slug": "ooak-data", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ad1b7cce12014a7746a7f8b912560717f6e2f037.png", "website": "https://ooakdata.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Ooak Data turns internal company data into a training ground where AI agents can be tested on real work.\r\n", "one_liner": "We turn company data into training data", "team_size": 5, "industry": "B2B", "subindustry": "B2B", "launched_at": 1783922301, "tags": [ "Reinforcement Learning", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/ooak-data", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/ooak-data.json", "year_founded": 2024, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/107463604", "x_url": "https://x.com/Ooakdata", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ad1b7cce12014a7746a7f8b912560717f6e2f037.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3207024, "full_name": "Pierre-Louis Vouteau", "title": "Founder/CEO", "founder_bio": "Cofounder & CEO at Ooak Data (YC S26). \nPrior COO @Visely (UK ed-tech)\nex Gopuff and L'Oréal", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/pierre-louis-vouteauau-15a06566/", "x_url": "https://x.com/PLouisV", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/465fc31f2866a593a25e59c4839739aafddbb167.jpg" }, { "user_id": 3207121, "full_name": "Thomas Aubry", "title": "Founder/CTO", "founder_bio": "Founder CTO at OoakData. \nPreviously Head of Data at PayLead, where he led the Data & AI team. Applied ML Scientist at Samsung AI. Earlier co-founded Macro Vision on autonomous vehicles Engineering. \nEducated at Institut supérieur de mécanique de Paris in Mechatronics and Robotics.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/thomaub", "x_url": "https://x.com/ThomAub", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1cb368f1fd2593cea6dfaa5e585238e725907fcc.jpg" } ], "jobs": [ { "id": 96721, "title": "ML engineer", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/M6anRxc-ml-engineer", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "€70K - €100K EUR", "equity_range": "0.10% - 0.30%", "experience": "6+ years", "visa": "US citizenship/visa not required", "skills": [ "Machine Learning", "Reinforcement learning (RL)" ] }, { "id": 99220, "title": "Software Engineer Senior ", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/KqZT9J1-software-engineer-senior", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "€55K - €80K EUR", "equity_range": "0.10% - 0.30%", "experience": "6+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 99221, "title": "Data engineer senior", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/0CZAefT-data-engineer-senior", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Engineering", "role_type": "Backend", "salary_range": "€65K - €95K EUR", "equity_range": "0.10% - 0.30%", "experience": "3+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 99325, "title": "Founding Marketing Lead ", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/atIhXqo-founding-marketing-lead", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Marketing", "role_type": null, "salary_range": "€50K - €80K EUR", "equity_range": "0.10% - 0.50%", "experience": "6+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 107967, "title": "1st Product Manager ", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/GqbqtyS-1st-product-manager", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Product", "role_type": null, "salary_range": "€70K - €90K EUR", "equity_range": "0.10% - 0.30%", "experience": "6+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 112051, "title": "Head of Engineering ", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/cFitSes-head-of-engineering", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "€100K - €130K EUR", "equity_range": "", "experience": "11+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 114401, "title": "Account Executive - Data Partnerships (US)", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/q01Cgyi-account-executive-data-partnerships-us", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "$100K - $140K", "equity_range": "", "experience": "6+ years", "visa": "US citizen/visa only", "skills": [] }, { "id": 114402, "title": "GTM Manager (US)", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/BJ8W0DT-gtm-manager-us", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "$120K - $150K", "equity_range": "", "experience": "6+ years", "visa": "US citizen/visa only", "skills": [] }, { "id": 114403, "title": "Head of Operations", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/fl4Oyho-head-of-operations", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "€80K - €100K EUR", "equity_range": "", "experience": "11+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 114404, "title": "Operations Manager - Data & AI", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/j3mCCMu-operations-manager-data-ai", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "€45K - €55K EUR", "equity_range": "", "experience": "1+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 114405, "title": "Data Quality Specialist", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/Gr7Gc1x-data-quality-specialist", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "€65K - €80K EUR", "equity_range": "", "experience": "6+ years", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 116458, "title": "Data Partnerships - Account Executive (EU)", "url": "https://www.ycombinator.com/companies/ooak-data/jobs/S7N4ejR-data-partnerships-account-executive-eu", "location": "Paris, IDF, FR / Paris, Île-de-France, FR", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "€55K - €70K EUR", "equity_range": "", "experience": "3+ years", "visa": "US citizenship/visa not required", "skills": [] } ], "news": [], "launches": [ { "id": 110403, "title": "Ooak Data: Turn internal company data into a training ground where AI agents can be trained on real work", "tagline": "The bottleneck to smarter agents isn't compute. It's real data. We unlock it.", "body": "**TL;DR** Ooak Data turns internal company data into a training ground where AI agents can be tested on real work.\\\n\\\n[https://drive.google.com/file/d/1EgbFzN-g4cVY-dk7ZPH6gR37F28D56xo](https://drive.google.com/file/d/1EgbFzN-g4cVY-dk7ZPH6gR37F28D56xo/view?usp=sharing)\n\n**The problem**\n\nAI agents still fail on real work: the messy, multi-tool, permission-heavy workflows inside actual companies. You can't synthesize it either, real org charts, permissions, and cross-tool dependencies don't come out of a generator. You have to source it from real companies and make it safe. That's the hard part, and it's what we do.\n\n**Our Solution**\n\nWe connect to a company's tools (Gmail, Slack, Notion, Jira, Drive, SharePoint) and anonymize everything into a digital twin: structurally identical, but privacy-safe. Labs use our product to train and evaluate AI agents on real work.\n\nWe're working with multiple frontier labs and are now well above 7 figures in revenue. We've cut ingestion-to-delivery time 3x since launch, across data from 20 companies.\n\n**The team**\n\nWe’re a team of 3 cofounders: Pierre-Louis, Grégoire and Thomas and spent our careers inside the messy company systems we now replicate.\n\n**Our asks**\n\n* If you're training or evaluating agents on real workflows, we'd love to talk.\n* Connections with companies abo 20+ FTEs, operating, winding down, or recently closed. Know someone who fits? **Add them [here](https://form.typeform.com/to/PjHdZmDW)**, or if it's your own company, grab [**15 min**](https://calendar.google.com/appointments/schedules/AcZssZ142jxQGEjB1PR9skAe17Nf7Pjf8G8EszSPcRhZIzmZXGCkn05REmmAmqPvMgWTYUMwbQNwinIF). Curious what a dataset is worth? Try our [**simulator**](https://ooakdata.com/partner).\\\n feel free to reach out to us: [pierrelouis@ooakdata.com](mailto:pierrelouis@ooakdata.com)\n\n![uploaded image](/media/?type=post&id=110403&key=user_uploads/3207024/bddd7844-32fc-4746-a3bd-e2a8450a85bd)\n\nYou can also reach us here: [pierrelouis@ooakdata.com](mailto:pierrelouis@ooakdata.com)", "url": "https://www.ycombinator.com/launches/Sih-ooak-data-turn-internal-company-data-into-a-training-ground-where-ai-agents-can-be-trained-on-real-work", "votes": 26, "created_at": "2026-08-14T15:00:00.284Z" } ] }, { "id": 32583, "name": "Graphify Labs", "slug": "graphify-labs", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/827fabd6e00bab1febd0feb00c906a6a3809b989.png", "website": "https://www.graphify.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Graphify is a Knowledge Graph control plane for enterprise software. The open-source tool has 116K+ GitHub stars and 6.5M+ downloads, and is used in production by engineers at Shopify, Datadog, JP Morgan, American Express, Harvey, Automation Anywhere, and Vanguard.\r\n\r\nOn top of that graph, our enterprise layer reviews every pull request and formally verifies each code change: it proves the new behavior matches the old, or hands back the exact input that breaks it. That makes it built for the changes teams fear most, like refactors, framework upgrades, and large-scale migrations, where you need proof that nothing broke, not a guess.\r\n\r\nThe graph updates itself as the code changes and keeps context instead of forgetting it, so your whole team and their coding agents work from one shared, always-current view of the codebase instead of a thousand private threads.\r\n\r\nIt runs entirely inside your perimeter, fully air-gapped on-premise or self-hosted in your own cloud, so no code or data ever leaves your environment.", "one_liner": "On-device Knowledge Graph engine for Enterprise Software", "team_size": 1, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1785106297, "tags": [ "Developer Tools", "Open Source" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/graphify-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/graphify-labs.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/graphify-labs", "x_url": "https://x.com/graphify", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/graphify-65db", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/827fabd6e00bab1febd0feb00c906a6a3809b989.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3674994, "full_name": "Safi Shamsi", "title": "Founder/CEO", "founder_bio": "Founder/CEO at Graphify Labs (YC S26). Built Graphify, an open-source knowledge graph engine that reached 116K+ GitHub stars and 6.5M+ PyPI downloads in about five months. Published researcher and author of The Memory Layer. MS Data Science, University of Birmingham; thesis on knowledge-graph-powered RAG systems achieving a 50% improvement in retrieval and 67% reduction in hallucinations.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/safi-shamsi/", "x_url": "https://x.com/safishamsii?s=21", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/49a0b9425549adb1bf790e96925ee78d020661c3.jpg" } ], "jobs": [], "news": [ { "title": "Your AI Agents Can Write Code. Can They Understand Your Business?", "url": "https://www.forbes.com/sites/stevenwolfepereira/2026/10/02/your-ai-agents-can-write-code-can-they-understand-your-business/", "date": "Oct 01, 2026" } ], "launches": [ { "id": 111176, "title": "Graphify Labs: Knowledge Graph Engine for Enterprises 🧠", "tagline": "Shared memory of your codebase, for humans and agents.", "body": "Hey everyone, I'm [Safi](https://www.linkedin.com/in/safi-shamsi/), founder and CEO of [Graphify Labs](https://www.linkedin.com/company/graphify-labs/).\n\nFor fifty years, the hard part of software was writing it. AI flipped that overnight. It writes more code than any team can read, and a lot of it is slop. That leaves two **hard problems**: no one can hold the whole codebase in their head, and no one can be sure an AI's change didn't quietly break something. **Graphify solves both.**\n\n**By the numbers**\n\n\\- **108K+ GitHub stars**: one of the top five open-source projects ever from a YC company, and the first to cross 100K⭐ mid-batch\n\n\\- **5M+ downloads** of the open-source engine\n\n\\- **6,000+ signups** on the platform in its first weeks\n\n\\- Enterprises already running it in production\n\n**The memory.** One shared, always-current knowledge graph of your whole codebase, across every language and every repo, plus its docs and tickets, so a change in one service shows its impact across all the others. Instead of grepping through files one at a time, your engineers and their AI agents trace how everything connects, who calls what and what a change touches, and answer questions across the entire system, at a **fraction of the tokens**. Agents connect over a single **MCP endpoint**, so the graph works inside Claude Code, Cursor, and every major coding tool. It groups the code into communities, the real subsystems, so teams can plan, onboard, and understand the architecture across the whole **SDLC**.\n\n**The proof.** Graphify reviews every pull request, including changes that span **multiple repositories**, and verifies each one. It writes and runs a test that reproduces the bug, proves the new code behaves exactly like the old, or hands back the exact input that breaks it. Most AI reviewers just have one model comment on another model's code, which is still a guess. **Graphify proves its findings**, because it runs on a real **code graph and formal verification**.\n\n![uploaded image](/media/?type=post&id=111176&key=user_uploads/3674994/e7f719ea-bca0-4c12-a922-12c728426dd4)\n\n**It learns.** Graphify remembers what your team ships and what it rejects, and feeds that back in, so its reviews get sharper over time and stop flagging what your team has already decided is fine.\n\n> Migrating a large codebase with AI is like changing a train's engine while it is moving. Graphify lets you do it without stopping the train, and proves the new version does exactly what the old one did.\n\n**Here's the launch film**:\n\n\\\n\\\nIt also proposes a fix and keeps it only if it re-verifies through the fix loop. Graphify connects to **Jira, Linear, and Sentry** so it knows the ticket a change should satisfy and the production bug it should resolve.\n\nThe enterprise layer runs on-prem, so your code never leaves your environment.\n\n**The founder.** [Safi Shamsi](https://www.linkedin.com/in/safi-shamsi/) is a published knowledge graph researcher and the author of The Memory Layer. His research on knowledge-graph retrieval improved retrieval 50% and cut hallucinations 67%. Now he's applying that work to one of the hardest problems in software: giving every engineer and AI agent a shared memory of the codebase, and a way to prove their changes still work.\n\nSoon every line of software will be written or reviewed by an agent. That software will need a memory of itself and proof that it works. **We're building both.**\n\n![uploaded image](/media/?type=post&id=111176&key=user_uploads/3674994/439244a1-338f-46cc-8594-b7c2a74bffcb)\n\n**The ask:** Before September 10 we're taking 10 design partners: a free 14-day build where we deploy the enterprise graph on-prem, with your team \\\n**Reach us at [founders@graphify.com](mailto:founders@graphify.com)**\n\n**Dig in**\n\n**- Run it yourself: [https://app.graphify.com](https://app.graphify.com)**\n\n**- Open-source engine: [https://github.com/Graphify-Labs/graphify](https://github.com/Graphify-Labs/graphify)**\n\n**- Docs: [https://graphify.com/docs](https://graphify.com/docs)**\n\n**- Connect any agent over MCP: [https://api.graphify.com/mcp](https://api.graphify.com/mcp)**\n\n**- Benchmarks and methodology: [https://github.com/Graphify-Labs/graphify/blob/v8/BENCHMARKS.md](https://github.com/Graphify-Labs/graphify/blob/v8/BENCHMARKS.md)**\n\n**- Talk to us: [founders@graphify.com](mailto:founders@graphify.com)**\n\n![uploaded image](/media/?type=post&id=111176&key=user_uploads/3674994/6068aaba-5948-422b-b5ef-b74b60a9504f)\n\n", "url": "https://www.ycombinator.com/launches/SvA-graphify-labs-knowledge-graph-engine-for-enterprises", "votes": 66, "created_at": "2026-08-19T15:00:00.223Z" } ] }, { "id": 32584, "name": "Aerogen Systems", "slug": "aerogen-systems", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/235b4dc8ef4a94e38d79ce892c9d55184a2b0224.png", "website": "https://www.aerogensystems.com/", "all_locations": "Ann Arbor, MI, USA", "long_description": "We are building the future of chip manufacturing.\r\n\r\nChips are more advanced than ever, and the features inside them keep getting smaller. But the cleanroom infrastructure supporting their manufacture has not scaled at the same pace. Intel, TSMC, Samsung, and others spend billions building and operating massive cleanrooms, only to isolate their most advanced wafers from the surrounding facility as much as possible.\r\n\r\nWe think that architecture is backwards. Instead of making the entire building cleaner, why not make the machinery itself as clean as possible?\r\n\r\nTalos is our proprietary tool-integrated purification system. It sits directly inside semiconductor equipment and continuously destroys airborne molecular contaminants before they reach the wafer. Instead of fighting contamination by constantly flooding tools with ultra-high-purity nitrogen, Talos keeps the environment inside the machine 10–100x cleaner, without the massive purge-gas bill.\r\n\r\nTalos is powerful on its own, but it is also the foundation for a much bigger vision: the Cleanpipe.\r\n\r\nThe Aerogen Systems Cleanpipe is an ultra-clean interconnect infrastructure that protects wafers as they move from machine to machine across the fab. Rather than relying on a billion-dollar cleanroom to protect every cubic foot of factory space, Cleanpipe creates a controlled environment only where cleanliness actually matters: around the wafer.\r\n\r\nTalos cleans the tools. Cleanpipe connects them. Eventually, we delete the cleanroom and build more advanced chips, at greater scale, with radically less infrastructure. With cleanpipes connecting tools, chip fabs can exist anywhere.", "one_liner": "Modernizing Chip Manufacturing Infrastructure", "team_size": 4, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1786900126, "tags": [ "Hard Tech", "B2B", "Manufacturing", "Semiconductors" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/aerogen-systems", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/aerogen-systems.json", "year_founded": 2026, "city": "Ann Arbor", "country": "US", "linkedin_url": "https://www.linkedin.com/company/aerogen-systems-company/", "x_url": "https://x.com/AerogenSystems", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/235b4dc8ef4a94e38d79ce892c9d55184a2b0224.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/e2f1aeb38c2211e11ef666f05de7b888fd393137.png" ], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 3388722, "full_name": "Noah Vogel", "title": "Founder/CEO", "founder_bio": "Founder & CEO at Aerogen Systems. Previously engineering intern at SpaceX building Starship actuators and at Draper developing advanced mission simulators. BSE Aerospace Engineering from University of Michigan.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/noahvogel/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5153adbbedb2cd7f09db195d1d63de5e64513dea.jpg" }, { "user_id": 3422652, "full_name": "Colin Lavery", "title": "Founder/CTO", "founder_bio": "Co-founder & CTO of Aerogen Systems. Previously an electrical engineer at H3D, developing 3D gamma ray imaging systems. BSE Space Sciences and Engineering at the University of Michigan.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/colin-lavery-67b7512a3/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3b3f944b8190b3fbc2ca61bd6963d1d3707c6976.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111026, "title": "Aerogen Systems: Building Chip Fabs Without Cleanrooms", "tagline": "Redefining the architecture behind chip manufacturing", "body": "Hi YC!\n\nWe’re [Noah](https://www.linkedin.com/in/noahvogel/) and [Colin](https://www.linkedin.com/in/colin-lavery-67b7512a3/), the founders of [Aerogen Systems](https://www.aerogensystems.com/). We’re building the next generation of clean infrastructure for chip manufacturing.\n\n**TL;DR:** Aerogen Systems makes semiconductor tools clean themselves. Talos, our tool-integrated purification system, continuously destroys airborne contaminants around the wafer, creating 10–100x cleaner tool environments while reducing dependence on ultra-high-purity nitrogen. Over time, Talos becomes the foundation for the Cleanpipe, an ultra-clean interconnect network that lets wafers move between tools without depending on the surrounding cleanroom.\n\n\n\n# **The Problem**\n\n**Chip features keep getting smaller, but cleanrooms do not scale with them.**\n\nSemiconductor manufacturers spend billions building enormous cleanroom facilities, then spend even more trying to isolate their most advanced wafers from the environment they just built.\n\nAs features approach the atomic scale, trace airborne contaminants that were once tolerable can create defects, reduce yield, and disrupt some of the most sensitive manufacturing processes on Earth. Yet the industry's answer has largely been more of the same: cleaner facilities, more filtration, and more ultra-high-purity nitrogen flooding sensitive equipment.\n\n**The proof is in the yield numbers.** 2026 reporting puts Intel’s 18A program between roughly 50% and 65% yield as it pushed the node toward high-volume production. With global semiconductor sales projected to reach **$1.5 trillion** in 2026, every percentage point matters. Improvements in yield and manufacturing efficiency can represent tens of billions of dollars in value.\n\n# **The Solution**\n\nAt Aerogen Systems, we think the industry is approaching this problem backwards.\n\n**The whole factory does not need to be a massive cleanroom.** Only the environment immediately surrounding the wafer does.\n\nTalos is our proprietary, tool-integrated purification system. It sits directly inside semiconductor manufacturing equipment and continuously recirculates and purifies the environment around the wafer. Instead of constantly replacing contaminated gas with more ultra-high-purity nitrogen, Talos destroys contaminants at the point of use, offering 10–100x cleaner tool environments while dramatically reducing dependence on purge gas and consumable filtration.\n\n![uploaded image](/media/?type=post&id=111026&key=user_uploads/3388722/a095a4e2-31a2-4307-b8f1-bca417cb27d7)\n\n_1. Rendering of standalone Talos unit._\n\nBut Talos is only part of the vision.\n\nWith tool-integrated purification units as a foundation, we build **the Cleanpipe: ultra-clean interconnect infrastructure** that protects wafers as they move from machine to machine across the fab. Talos cleans the tools. The Cleanpipe connects tools to each other. Together, they delete the cleanroom and all the expenses that come with it.\n\n# **What We’ve Built**\n\nTalos has progressed through multiple generations of hardware development and is protected by patent-pending IP. We’ve validated the core purification architecture and are now working toward tool integrations, pilots, and commercial deployments with semiconductor manufacturers and equipment companies.\n\n![uploaded image](/media/?type=post&id=111026&key=user_uploads/3388722/a8596c2c-c29b-4271-8d92-85672dce0dc1)\n\n_2. Test results showing Talos bringing industry-standard 6N purification levels down to 8N — a 100x improvement in cleanliness._\n\nThe next step is connecting these purified environments together through the Cleanpipe, creating an ultra-clean path around the wafer from tool to tool and ultimately removing the need to maintain the entire fab as one enormous cleanroom.\n\n# **The Team**\n\nAerogen Systems started as a NASA-backed research group at the University of Michigan, where our team developed the underlying purification technology. \n\nWe have spent the last several years designing, testing, patenting, and rebuilding this technology for extreme manufacturing environments.\n\n* Noah, CEO, worked in manufacturing engineering at SpaceX, where he focused on production rate and yield improvement for the Starship program.\n\n\n* Colin, CTO, previously built advanced particle detectors at H3D and led the team through the earliest design cycles of Talos.\n\n![uploaded image](/media/?type=post&id=111026&key=user_uploads/3388722/4db2a255-30a8-44ee-928b-b07425918094)\n\n# **Our Ask**\n\nWe want to meet semiconductor fabs, equipment OEMs, and manufacturing teams interested in pilots, tool integrations, and commercial deployments. If this sounds like you, please reach out to [hi@aerogensystems.com](mailto:hi@aerogensystems.com)!\n\nWe are building toward a world where fabs no longer need billion-dollar cleanrooms to manufacture the world’s most advanced chips.", "url": "https://www.ycombinator.com/launches/Ssk-aerogen-systems-building-chip-fabs-without-cleanrooms", "votes": 17, "created_at": "2026-08-18T21:15:05.234Z" } ] }, { "id": 32585, "name": "Hilstart", "slug": "hilstart", "former_names": [ "HILStart", "HILstart" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/51bcabfa16bc955e9ce96986bcffa07d33aea4b7.png", "website": "https://hilstart.io/", "all_locations": "San Francisco, CA, USA", "long_description": "We build physical agent harnesses to enable AI to debug, validate, and ship functional hardware and software.", "one_liner": "Autonomous Hardware Testing", "team_size": 0, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1787080990, "tags": [ "Developer Tools", "Hardware" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/hilstart", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/hilstart.json", "year_founded": null, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/51bcabfa16bc955e9ce96986bcffa07d33aea4b7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3650587, "full_name": "William A'court", "title": "Founder/CEO", "founder_bio": "Founder and CEO at Hilstart. \n\nPreviously Lead Mechatronics Engineer at Aquila, shooting lasers at drones to charge them.\n\nAlso made FSAE race cars and electric jetboats along the way.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/william-acourt", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1ed4a0dd114a57e9fad794d4d76ca2b92b744087.jpg" }, { "user_id": 3673236, "full_name": "Simon Alexander", "title": "Founder/CTO", "founder_bio": "Founder & CTO at Hilstart. \n\nPreviously Firmware Engineer at Aquila.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/simon-alexander-8045a685/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f81fdbde1bd9795d0673493281c06b35e6e75e71.jpg" }, { "user_id": 3609313, "full_name": "Rahul Bhati", "title": "Founder", "founder_bio": "Trained gymnast. Building hardware since 15. 2x National robotics champion at 18. \n\nBefore HILStart: power and thermal systems at Vessev (Flying boats) and test rigs at Halter (NZ's latest unicorn). Founding team at Aquila and built the core tech behind kilometre-scale wireless laser energy transmission across many engineering disciplines. Aquila is now a world-leading powerbeaming company, and the tech delivered the world's first 500W transfer to a drone mid-flight.\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rahulbhati240p/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/dcade1c338d4f7cd84c76029a88c17e2378d7677.jpg" }, { "user_id": 4019599, "full_name": "Josh Schoeman", "title": "Founder", "founder_bio": "Making hardware easier.\nPast adventures all over mechanical and electrical development.\n\nBattle scars from electric motor design, accelerated lifecycle testing, mechanical structures, cooling systems, hardware test rigs, DFM, wiring harnesses, race cars and electronics debugging. \nAlso, destructive testing (often inadvertent).", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/joshua-schoeman-167a5b211/", "x_url": "https://x.com/Formula_Josh", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5b42ede95509dba6974a14a6bb16272b6d7579a8.jpg" } ], "jobs": [], "news": [], "launches": [] }, { "id": 32586, "name": "Greypoint Industries", "slug": "greypoint-industries", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/b0d5898f7b8fc2ee6fc19f6de3dae045123c7608.png", "website": "https://greypointindustries.com/", "all_locations": "San Francisco, CA, USA", "long_description": "We build LEGION, a drone swarm that identifies, geolocates, and tracks every emitter on the battlefield using the enemy’s own radio signals. LEGION enables our warfighters to precisely track previously obscured targets like jammers, command posts, and drone operators so they can proactively respond to threats instead of reacting to them.", "one_liner": "Building drone swarms that hunt hidden enemies.", "team_size": 4, "industry": "Industrials", "subindustry": "Industrials -> Defense", "launched_at": 1782765221, "tags": [ "Hardware", "Robotics", "Drones", "Defense" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Defense" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/greypoint-industries", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/greypoint-industries.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/greypoint-industries/?viewAsMember=true", "x_url": "https://x.com/Greypoint_", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/b0d5898f7b8fc2ee6fc19f6de3dae045123c7608.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 3676390, "full_name": "Darcy Zhang", "title": "Founder/CEO", "founder_bio": "Co-Founder at Greypoint Industries.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/darcy-zhang-021134243/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5337a388dc0b683be08ad652cef050e1fd9af004.jpg" }, { "user_id": 3676391, "full_name": "Nic Chu", "title": "Founder/COO", "founder_bio": "Co-Founder at Greypoint Industries.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/nicchu/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2c6aa1d0091b93ae6f6641c720d544f1c7126a60.jpg" }, { "user_id": 3558298, "full_name": "Jordan Khoo", "title": "Founder/CTO", "founder_bio": "Co-Founder at Greypoint Industries. Previously led project with autonomous loitering munition project with Lockheed Martin Skunkworks.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jordankh00/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2363824c9001ce22d79d36c04d0ab6555b86edfa.jpg" }, { "user_id": 3676397, "full_name": "Steven Xu", "title": "Founder/CPO", "founder_bio": "Co-Founder at Greypoint Industries. Previously at Bosch Quantum Sensing building satellite sensors. Led research project on detection of UAVs using passive radar. Presidential Scholar at UBC.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/sxu22/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9cffa849cc769fd0df23fe0a5e3b4c24fb8e7736.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108367, "title": "Greypoint Industries: Drone swarms that find drone operators", "tagline": "Greypoint Industries builds drone swarms that find enemy drone operators on the battlefield.", "body": "[Greypoint Industries (YC S26) Launch Video](https://youtu.be/y9jC4A7Fg04)\n\nGreypoint Industries builds drone swarms that find enemy drone operators on the battlefield.\n\nOver 70% of casualties in Ukraine come from drones. The entire defense space is focused on building interceptors to shoot them down. While they’re treating the symptom, we’re going after the root cause: the drone stations. They’re hidden in bunkers, behind terrain, and fly the drones that kill our warfighters and civilians. Today, soldiers have no way to reliably geolocate operators on the battlefield.\n\nGreypoint Industries built LEGION, a drone swarm that identifies, geolocates, and tracks every emitter on the battlefield using the enemy’s own radio signals. LEGION enables our warfighters to precisely track previously obscured targets like jammers, command posts, and drone operators so they can proactively respond to threats instead of reacting to them.\n\nGreypoint has just won a contract with the Canadian Government, demoed with the Armed Forces, and successfully geolocated a target using their aerial platform. The founding team brings experience from the Canadian Infantry, General Dynamics, Bosch Quantum Sensing, and TerraSense Analytics.\n\nDefend the West. Deploy Greypoint.\\\n\\\nFind out more at .", "url": "https://www.ycombinator.com/launches/SBr-greypoint-industries-drone-swarms-that-find-drone-operators", "votes": 10, "created_at": "2026-07-30T23:00:42.765Z" } ] }, { "id": 32587, "name": "TareBio", "slug": "tarebio", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/9e770bbea41518001119968d389d0d3e35ce7abe.png", "website": "https://tarebio.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Our platform develops personalized or population-level vaccines for any disease in weeks by changing only one input: the immune imprint. Instead of trying to predict which antigens might work, we read out immune cells already fighting the disease in the body and reverse-engineer antigens that activate them. We've already run our platform end-to-end and created a vaccine that cured cancer in mice.", "one_liner": "Vaccines for any cancer or infection in weeks.", "team_size": 1, "industry": "Healthcare", "subindustry": "Healthcare -> Drug Discovery and Delivery", "launched_at": 1785462754, "tags": [ "Vaccines", "Synthetic Biology", "Biotechnology" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare", "Drug Discovery and Delivery" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/tarebio", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/tarebio.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/tarebio/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/9e770bbea41518001119968d389d0d3e35ce7abe.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/96525419f2b8f6b4da09d8c897eb1c8553bef26f.png" ], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 431020, "full_name": "Eta Atolia", "title": "Founder/CEO", "founder_bio": "Founder & CEO, TareBio (YC S26). MD-PhD from UCLA-Caltech, where she built a novel platform that develops vaccines for any cancer or infection in weeks, funded by the Gates Foundation. BS in Biology from MIT. Conducted research at the Koch Institute, Harvard Medical School, and the Weizmann.\n\nSelect Awards: UCLA MD-PhD Merit Scholar, UCLA TDG Senior Fellow, MIT iGEM Gold Medal, MIT IDEAs winner, MIT Dean Advisory Group, WhiteHouse Science Fair, Bessie Lawrence Scholar, Intel STS Finalist", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/eta-atolia/", "x_url": "https://x.com/eatolia", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1b89afa4d7ba0012222ebc57912ad8e652db40bd.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108804, "title": "TareBio: Vaccines for any cancer or infection in weeks", "tagline": "We make personalized or population-level vaccines reverse-engineered from the immune system. We have already cured cancer in mice and are vaccinating our first patient within the year.", "body": "**TL;DR:** Our platform develops **personalized** or **population-level** vaccines for **_any_ disease** in weeks by **changing _only one input_**: the immune imprint. Instead of trying to predict which antigens might work, we read out immune cells already fighting the disease in the body and reverse-engineer antigens that activate them. We've already run our platform end-to-end and created a vaccine that cured cancer in mice. We will be vaccinating our first patient within a year.\n\n![uploaded image](/media/?type=post&id=108804&key=user_uploads/431020/46cdbd4d-72d9-4231-a781-01a886dd82a4)\n\nView our launch video here: \n\n**The vaccine bottleneck**\n\nFor 200+ years, vaccines have been designed using the same workflow: study a pathogen or tumor, _predict_ which antigens might trigger protection, vaccinate, and hopefully generate a good immune response. This works for simpler infections. It has struggled for decades on complex diseases like HIV, TB, Hep C, and nearly _every cancer_. Currently, vaccine design is treated as a prediction problem.\n\n**What we do differently**\n\nInstead, we treat it as a readout problem. The **immune system has _already solved_** antigen selection in the body, it knows what it's responding to and the best immune cells for the job. So, we flip the whole process around:\n\n1. Capture the immune cells already responding to a cancer or infection, the \"**immune imprint**\"\n2. Use this imprint to directly identify, or “**reverse engineer**”, activating antigens\n3. Build a **vaccine** from those antigens\n\n![uploaded image](/media/?type=post&id=108804&key=user_uploads/431020/31744f0e-4a44-4aa9-aa2e-29d174a8d30a)\n\nThe same platform works for **_any_ disease** by changing one input: the immune imprint. Function over prediction, no constrained search space, and the output is a vaccine that can generate durable immunity in the body, not a transient antibody conjugate or cellular therapy.\n\n**Proof**\n\nWe built the full platform in mice and ran it end-to-end: identified antigens from the immune imprint of the cancer, designed a vaccine from them, and **cured lymphoma in mice.** Nothing about the technology is specific to lymphoma, it generalizes to any cancer or infection. We are using our technology to create both personalized (n=1) and population-level vaccines.\n\n**Why us**\n\nI'm Eta Atolia, founder and CEO. I built this platform over the last 8 years as part of my MD-PhD at UCLA-Caltech with funding from The Gates Foundation. Before this, I did my undergrad at MIT with most of my time dedicated to cancer immunotherapy and infectious disease research at the Koch and HMS. Our scientific advisory board includes Darrell Irvine, PhD (Vice Chair of Immunology at Scripps, Former MIT Professor of 22 years, world-renowned vaccine engineer), Manish Butte, MD, PhD (Chief of Immunology at UCLA and expert rare disease clinician), and John Glaspy, MD, MPH (Endowed Chair and Medical Director of Integrative Oncology at UCLA).\n\n**Our ask**\n\nWe are actively working with organizations and individuals to create both population-level vaccines and personalized treatments.\n\nHelpful things as we build:\n\n* If you have **patient tumor** **or** **blood samples**, please reach out.\n* Intros to **biotech + pharma** and **vaccine teams** for partnerships.\n* **Immunologists, oncologists, infectious disease specialists,** and **ML engineers** who want to build this with us.\n\nIf any of this resonates, reach out at [contact@tarebio.com](mailto:contact@tarebio.com).\n\n![uploaded image](/media/?type=post&id=108804&key=user_uploads/431020/e32e6406-ca9a-4740-9a77-48dd31f02e7f)\n\n", "url": "https://www.ycombinator.com/launches/SIu-tarebio-vaccines-for-any-cancer-or-infection-in-weeks", "votes": 13, "created_at": "2026-08-04T14:27:01.936Z" } ] }, { "id": 32588, "name": "Almanac", "slug": "almanac", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/04034530613a233afe0d28a52bcfd6e9943b7ea7.png", "website": "https://usealmanac.com", "all_locations": "San Francisco, CA, USA", "long_description": "Almanac is an AI agent that deeply understands your work.\r\n\r\nToday people rely on a stack of productivity tools to hold the context of their work including todo lists, issue trackers, wikis, and CRMs. All of them have to be updated by hand, and none were built for AI. \r\n\r\nAlmanac maintains its own self-updating productivity suite for you. This gives it an absurdly good understanding of your work, based on which it proactively gets things done for you.\r\n\r\nAlmanac starts with individuals, but it's built for teams. As more of a team connects their work, the agent gains shared context across everyone, so it understands not just what you're doing but how the whole team's work fits together.", "one_liner": "Organizing company knowledge", "team_size": 3, "industry": "B2B", "subindustry": "B2B", "launched_at": 1784235632, "tags": [ "Generative AI", "AI", "AI Assistant" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/almanac", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/almanac.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/usealmanac/", "x_url": "https://x.com/usealmanac", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/04034530613a233afe0d28a52bcfd6e9943b7ea7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 2550783, "full_name": "Rohan Sharma", "title": "Founder/CEO", "founder_bio": "Cofounder at Almanac. ML research at the Vector Institute, CMU and MIT. MEng in Computational Science from Harvard, BTech from IIT Delhi.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rohan-sharma-b6275219b/", "x_url": "https://x.com/rohansharma0509", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a414112b79dc50429aa2b090b2c58b1a3114d38c.jpg" }, { "user_id": 2557429, "full_name": "Kushagra Chitkara", "title": "Founder/COO", "founder_bio": "Cofounder at Almanac. Masters in Data Science from Harvard '25, Bachelors in Electrical from IIT Kharagpur '23.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/kushagra-chitkara-2b3226192", "x_url": "https://x.com/kushagrchitkar", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a782605dfc529d9f7a47e2f0399e2cf9cf525470.jpg" }, { "user_id": 2251543, "full_name": "Divit Sheth", "title": "Founder/CTO", "founder_bio": "Cofounder at Almanac. Previously Senior Software Engineer at Arcesium. BE in Electrical and Electronics Engineering from BITS Pilani with a minor in Data Science.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/divit-sheth/", "x_url": "https://x.com/divitsheth007", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/93674e13d8af5f7bcc49866d253a7ccba619b8d4.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 106258, "title": "Almanac: The agent with a company brain", "tagline": "Almanac works with its own browser and terminal, signed into the accounts you connect.", "body": "There's a new kind of agent.\n\nIt exists on its own computer, connected to your tools, with a living brain of everything you and your company know.\n\nIntroducing Almanac, the agent with a company brain.\n\n\n\nAlmanac works with its own browser and terminal, signed into the accounts you connect. Everything in your tools compiles into a wiki that it reads before every task. Your teammates' Almanacs share that wiki. You text it work. It texts you when it's done.\n\n![uploaded image](/media/?type=post&id=106258&key=user_uploads/2550783/615174e2-3d24-480d-a8f5-7c849c5a24b9)\n\n**What we hand ours**\n\n* Divit has it gather feedback from our customer Slack channels and file GitHub issues, with screenshots and reproduction steps.\n* Kushagra connected his Chase account. It inspects recurring charges and cancels subscriptions he doesn't use.\n* Before meetings, it sends us a brief built from our past emails and calendar. After, it drafts the follow-up. The draft is usually good enough to send.\n* It watches our production environment. When something trivial breaks, it fixes it and raises a PR.\n* It buys our dinner. At this point it suggests the order.\n\n**It knows when to stop.** At a login, a payment, or a decision it shouldn't make alone, it pings you or hands you the live browser.\n\n![uploaded image](/media/?type=post&id=106258&key=user_uploads/2550783/9b85821d-c37b-4da7-87a9-6218d731ad53)\n\n**Automations are just sentences.** \"Whenever a customer reports a bug, check whether anyone else has reported it, reproduce it, and update the GitHub issue.\" Say it once. There's no workflow builder.\n\n**The brain is shared.** Your teammates' Almanacs read and write the same wiki. So can your other agents: Claude Code and Codex can start caught up on your company. \\\n\\\n**Our bet.** We're Rohan, Kushagra, and Divit. We grew up together in India and met 11 years ago preparing for the JEE. Since then: machine-learning research and engineering across Vector Institute, CMU, MIT, Harvard, and Arcesium. From the start we've had one bet: AI will need a context layer, and it takes the shape of a wiki. For the last year and a half we've built agent wikis for Harvard, for NASA, and once, an attempt at the whole internet. We are, perhaps predictably, obsessed with wikis. Almanac is that bet, pointed at your company.\n\n![uploaded image](/media/?type=post&id=106258&key=user_uploads/2550783/7bab1366-3e4c-4140-829d-e1bc9666d9ab)\n\nText it your first task tonight. See what's done by morning.\\\n\\\nTry it out today: ", "url": "https://www.ycombinator.com/launches/Rdq-almanac-the-agent-with-a-company-brain", "votes": 36, "created_at": "2026-08-13T15:37:10.813Z" } ] }, { "id": 32617, "name": "Whitespace", "slug": "whitespace", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/d7ab9b902afa3fe0a37b3622022557a9cd82d082.png", "website": "https://whitespacehq.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Whitespace builds AI agents that run the entire operation of supply chain businesses, starting with wholesale distributors. We give them self-improving AI agents that learn their business and do work like inventory planning and customer service.\r\n\r\nWe're building towards a world where all supply chain operations run through our agents, and all physical goods purchases can happen across businesses without human touch.", "one_liner": "AI Operating System for Wholesale Distribution", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Supply Chain and Logistics", "launched_at": 1782389037, "tags": [ "SaaS", "B2B", "Supply Chain", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Supply Chain and Logistics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/whitespace", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/whitespace.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/whitespaceai", "x_url": "https://x.com/whitespacehqai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/d7ab9b902afa3fe0a37b3622022557a9cd82d082.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 2335228, "full_name": "Alex Tung", "title": "Founder/CEO", "founder_bio": "Founder at Whitespace. First non-engineer hire at Ankar (Index-backed), scaling it to 7-figure ARR in under a year. Previously BCG, deploying AI at F500 distributors. Cambridge Economics.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/alexyctung/", "x_url": "https://x.com/alexytung", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/cf7ed849f4f96987aa9557ffb71c9a5f4c714a48.jpg" }, { "user_id": 3405544, "full_name": "Leon Yao", "title": "Founder/CTO", "founder_bio": "Founder at Whitespace. Previously Research Engineer at InstaDeep, running 7-figure supply chain AI deployments for F500 enterprises. UCL Machine Learning, publishing AI papers at IROS (top-3 robotics conference) as an undergrad.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/linghong-yao/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ef2cb473c4b76b2f791ba1cf0b953446391daa5e.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 105676, "title": "Whitespace: AI Operating System for Wholesale Distributors", "tagline": "We help wholesale distributors grow 100x with AI agents that learn their business and do the work.", "body": "Hey everyone 👋 We’re [Alex](https://www.linkedin.com/in/alexyctung/) and [Leon](https://www.linkedin.com/in/linghong-yao/) - cofounders of [Whitespace](https://whitespacehq.ai/).\\\n\\\nWhitespace is the AI operating system for wholesale distributors. It connects to their systems to build a living model of their business, then deploys self-improving AI agents to do work like inventory planning, customer service, and order processing.\n\n**Problem**\n\nWholesale distribution is a $60 trillion global industry that touches one in every five dollars of global output. Yet it still runs largely manually. At a typical wholesale distributor, 30-50% of staff are dedicated to back-office tasks.\n\nThey spend their days on email, spreadsheets, and legacy systems, relying on tribal knowledge in their heads. Businesses cannot scale without hiring, and training a new hire takes >6 months.\n\nA quote from one of our customers really brings this to life: “I have 6 reps typing 600 lines of orders from email into the ERP every day. Everyone is inside our ERP and spreadsheets, copying data from one place to another.”\n\n**Solution**\n\nWhitespace is the AI operating system for wholesale distributors. It connects to their systems to build a living model of their business, then deploys self-improving AI agents. \n\nWe’re already working with wholesale distributors in sectors including industrial supplies and building materials to set up agents for processes like:\n\n* **Inventory Planning.** Forecasts product demand, recommends reorder quantities against stock levels, and learns over time from actual demand.\n* **Customer Service.** Monitors inboxes for support queries, digs through systems to find answers, and responds with human oversight.\n* **Order Processing.** Monitors inboxes, enters orders into the ERP, flags exceptions to humans, and learns from their corrections.\n\n\n\n**Team**\n\nBoth our families ran wholesale distribution businesses, so we grew up seeing the inefficiencies and the opportunities firsthand.\n\n[Alex](https://www.linkedin.com/in/alexyctung/) advised F500 manufacturers and distributors on tech/AI at BCG, took a startup from launch to 7-figure ARR in a year as its first non-engineering hire, and graduated top of his college at Cambridge.\n\n[Leon](https://www.linkedin.com/in/linghong-yao/) built AI products for F500 logistics distribution companies at InstaDeep, published first-author RL papers at tier-1 conferences, and previously founded an EdTech startup that served thousands of users worldwide.\n\nWe’ve been best friends for 12 years since high school. We’ve lived and worked together, and are now tackling a challenge close to our hearts.\n\n![uploaded image](/media/?type=post&id=105676&key=user_uploads/2335228/2d187648-e6e7-4b98-8101-407f5e7aefee)\n\n**Ask**\n\nWe would appreciate intros to wholesalers or distributors at [alex@whitespacehq.ai](mailto:alex@whitespacehq.ai)", "url": "https://www.ycombinator.com/launches/RUS-whitespace-ai-operating-system-for-wholesale-distributors", "votes": 11, "created_at": "2026-07-16T14:54:16.399Z" } ] }, { "id": 32618, "name": "Click", "slug": "click", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/3eb675c91f3464d134cd3d9c18712ad08bea23bb.png", "website": "https://www.useclick.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Click serves agents over an API. Bring your model keys, tools, and context - Click runs the agent and streams back events.\r\nAsk your coding agent about useclick.ai/reference!", "one_liner": "Managed Agents API", "team_size": 1, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1785963586, "tags": [ "API", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/click", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/click.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/use-click", "x_url": "https://x.com/useclick_ai", "fb_url": null, "cb_url": null, "github_url": "https://github.com/use-click/", "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/3eb675c91f3464d134cd3d9c18712ad08bea23bb.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/29bbb33ca2f50ca76c9e6e410f2b4252a484c4e9.png" ], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 1995988, "full_name": "Aditya Asgaonkar", "title": "Founder/CEO", "founder_bio": "Founder at Click. Previously helped build Ethereum's distributed system (proof-of-stake) as an open-source contributor at the Ethereum Foundation.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aditya-asgaonkar/", "x_url": "https://x.com/adiasg", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1d21221f7711a082e5615e10bb6d4053d1be4205.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110141, "title": "Click - Research services inside Codex & Claude", "tagline": "Live external context from professional & social platforms, marketplaces, financials, and more.", "body": "**TLDR:** Click is an MCP that provides extensive research connectors to give external context that the built-in web search misses.\n\n![uploaded image](/media/?type=post&id=110141&key=user_uploads/1995988/91aff538-41b1-4c8b-96eb-c1bf84bd5e47)\n\n**Video:** \n\n**Problem:**\n\nClick is largely motivated by a change in my own behavior:\n\nCodex & Claude are amazing products. I live inside their surface now - using it more than my browser. It feels like home, to the point that I’m annoyed to have to leave. Most often, I’m stepping outside to do external-facing tasks, like researching my leads on LinkedIn or comparing flight options. I want my agents to do these tasks as well.\n\nCurrently, those can’t be completed because their gateway to the world (in-built web search) has limited access. Often the agent also asks to use the browser with a logged-in session, which I’m uncomfortable with.\n\nSo, I built the connector layer that I wish existed.\n\n**Solution:**\n\nClick’s MCP is a single installation that provides your agent with an ever-growing bundle of the best data connectors. The current set of connectors enable new use cases inside the chat, like:\n\n* Lead research & enrichment\n* Researching ad campaigns from competitors\n* Analyzing sentiment around a topic across social media\n* Planning travel with live fares & availability\n\n**Team:**\n\nI’m Aditya, the founder of Click. Previously, I helped build Ethereum, with a vision to create a platform that enables new kinds of applications. \n\nNow, I’m focusing on the next application platform - the AI chat surface. Click builds useful services that meet users right inside the conversation. I’m building one trusted service at a time, and research is the first one.", "url": "https://www.ycombinator.com/launches/SeT-click-research-services-inside-codex-claude", "votes": 10, "created_at": "2026-08-12T19:00:01.957Z" } ] }, { "id": 32619, "name": "Akon Labs", "slug": "akon-labs", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/7195cae7a2ed2e9a5da8741119043c95ba614f7c.png", "website": "https://akonlabs.com", "all_locations": "San Francisco, CA, USA", "long_description": "", "one_liner": "Databricks for Enterprise Codebases", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1786669281, "tags": [ "Generative AI", "SaaS", "B2B", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/akon-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/akon-labs.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/akon-labs/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/7195cae7a2ed2e9a5da8741119043c95ba614f7c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 1607382, "full_name": "Subham Kundu", "title": "Founder/CEO", "founder_bio": "Founder at Akon Labs (YC S26). Previously CTO at Cignara (YC P26), building from first line to millionth. Principal AI Engineer at HTCD, developed AI observability platform and reduced AI search costs by 78%. Founding Engineer at Caravel Labs, built LLM chatbot scaled to 5,000 students and featured on Chinese national TV.", "is_active": true, "linkedin_url": "https://linkedin.com/in/subham-kundu-2746b515b", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/95a2cbce998042d28b021894a4de3a4084143d54.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110659, "title": "Akon Labs: We're building the most token-efficient coding agent in the world", "tagline": "GitNexus gives coding agents a map of your codebase, so they stop burning tokens looking for it.", "body": "##### We're building the most token-efficient coding agent in the world. The reason it works is GitNexus, our open-source knowledge graph for your codebase.\n\n\n\n##### The problem\n\nCoding agents don't understand your repo. They grep, they guess, they re-read the same files, and they stuff half your codebase into a context window hoping the answer is in there somewhere. You pay for every one of those tokens, and the agent still misses the cross-file relationship that mattered.\n\nThe bottleneck isn't the model. It's context.\n\n##### What we built:\n\nGitNexus indexes your repository into a knowledge graph (symbols, call paths, imports, ownership, history) and serves the agent exactly the context a task needs, instead of everything it might need.\n\nIt's one connected stack:\n\n* The knowledge graph. Run npx gitnexus analyze and you have it\n* A coding agent that runs on top of it\n* Automated PR review grounded in the same graph\n\n![uploaded image](/media/?type=post&id=110659&key=user_uploads/1607382/1d3e51af-d3b5-4ee6-b0c3-b80bc4337c3e)\n\n##### The proof:\n\nOn the DeepSWE benchmark, we ran the same agent with our graph and without it:\n\n* 37% to 68.4% pass rate. Nearly twice the bugs solved\n* Half the cost per fix, with fewer tokens\n* On the hardest tasks, the gap more than triples\n\nFull methodology and results: \n\n##### Where we are:\n\n* 45,000+ GitHub stars\n* 65,000 weekly npm downloads\n* 10+ enterprises onboarded\n\n##### About the team:\n\nSubham Kundu and Abhigyan Patwari have spent nearly half a decade working on knowledge graphs. Between them, they've been founding engineers, CTOs, and principal AI engineers, taking products from 0 to 100 at multiple startups. Subham is a seasoned engineer who has scaled products to millions of users, won multiple hackathons, and doing a PhD in reinforcement learning. Abhigyan is a seasoned technologist and serial open source contributor, with involvement across multiple open source projects.\n\n##### Run it in your way:\n\nManaged, or fully sovereign. Self-hosted, air-gapped, nothing leaves your network.\n\n##### Looking to cut your coding agent bill?\n\nIf you want to lower your coding agent spend without giving up frontier performance, sign up at . We'll get you set up on your own repo and show you the before-and-after on your actual token costs.\n\n##### Our other asks\n\n* Try the open-source version. Run npx gitnexus analyze. Takes a minute\n* Intros. Engineering leaders where agent cost or code-context quality is a real line item, especially in regulated or air-gapped environments\n* Tell us where it breaks. Weird monorepos and polyglot repos are the most useful feedback we get\n\nGitNexus, by Akon Labs. The nervous system for agent context.\n\nContact us at: [founders@akonlabs.com](mailto:founders@akonlabs.com)\n\nWebsite: [https://www.akonlabs.com](https://www.akonlabs.com/)\n\nLinkedin: [https://www.linkedin.com/company/akon-labs](https://www.linkedin.com/company/akon-labs/)", "url": "https://www.ycombinator.com/launches/Smp-akon-labs-we-re-building-the-most-token-efficient-coding-agent-in-the-world", "votes": 45, "created_at": "2026-08-17T19:30:00.167Z" } ] }, { "id": 32621, "name": "Aktoria Robotics", "slug": "aktoria-robotics", "former_names": [ "Phi Robotics" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/87a10d62adb927072657a9be3281af92a78ab9c6.png", "website": "https://aktoria.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Enabling robot deployment & autonomy ", "one_liner": "Human as API for robots", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1784175637, "tags": [ "Robotics", "Enterprise Software" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/aktoria-robotics", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/aktoria-robotics.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/aktoria-robotics", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/87a10d62adb927072657a9be3281af92a78ab9c6.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3108880, "full_name": "Jaskaran Singh Walia", "title": "Founder", "founder_bio": "Ex - Researcher @ (Microsoft, MIT, CMU, Cambridge, NUS)\n\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jaskaranwalia", "x_url": "https://x.com/karanwxliaa", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/bb7e6a357374ed42415aee6999d0568b1d2dc6ce.jpg" }, { "user_id": 1125677, "full_name": "Shreenabh Agrawal", "title": "Founder", "founder_bio": "Robotics grad @ CMU. Physics & Quantum Technology @ IISc Bangalore.\nIndia's Prime Minister awardee, highest civilian honor for young achievers.\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shreenabh/", "x_url": "https://x.com/ShreenabhA", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/64c61d6ba30010f5699e5b194f9a5ff2aa45b8ef.jpg" } ], "jobs": [ { "id": 109688, "title": "Robot Learning Researcher (VLA / Foundation Models)", "url": "https://www.ycombinator.com/companies/aktoria-robotics/jobs/WlROHB1-robot-learning-researcher-vla-foundation-models", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Science", "role_type": "Research", "salary_range": "$450K - $600K", "equity_range": "1.00% - 2.00%", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [] }, { "id": 116148, "title": "HR Manager (India)", "url": "https://www.ycombinator.com/companies/aktoria-robotics/jobs/sIDJZkE-hr-manager-india", "location": "Noida, UP, IN / Noida, Uttar Pradesh, IN", "type": "Full-time", "role": "Recruiting & HR", "role_type": "Operations", "salary_range": "₹700K - ₹1.4M INR", "equity_range": "", "experience": "6+ years", "visa": "US citizenship/visa not required", "skills": [] } ], "news": [], "launches": [ { "id": 109426, "title": "Aktoria Robotics is the teleoperation layer to speed up robot deployments", "tagline": "A 24-hour global network of trained operators, powered by our low-latency teleop software.", "body": "**Aktoria Robotics is the teleoperation layer that speeds up robot deployments.**\n\n[https://drive.google.com/file/d/1UQchwD0U_FFdP9X3FDKJ7UjqXRW_zh24](https://drive.google.com/file/d/1UQchwD0U_FFdP9X3FDKJ7UjqXRW_zh24/view?usp=sharing)\n\nEvery \"autonomous\" robot you've seen this year had a human behind it. Full autonomy isn't here yet, and building an in-house teleop team is slow and expensive. That's the bottleneck standing between a working robot and a real deployment.\n\nAktoria Robotics is a 24-hour global network of trained robot operators across India, Mexico, and the USA, powered by our low-latency teleop software (40ms cross-country) at at least 50% less than the cost of in-house operators. Robotics companies plug in and ship deployments faster.\n\n**The team:** We're Jaskaran Singh Walia and Shreenabh Agrawal, ML and Robotics grad students at CMU. Shreenabh (ex-Zoox) has first-author papers at top robotics conferences. Jaskaran has authored 15+ papers and researched at MIT, Cambridge, and Microsoft Research.\n\n**Our ask:** If you're deploying robots and need reliable teleop, book a call through aktoria.com. Intros to robotics teams requiring teleop are hugely helpful, reply here or reach us at [founders@aktoria.com](mailto:founders@aktoria.com).", "url": "https://www.ycombinator.com/launches/SSw-aktoria-robotics-is-the-teleoperation-layer-to-speed-up-robot-deployments", "votes": 6, "created_at": "2026-08-06T15:00:00.066Z" } ] }, { "id": 32622, "name": "Rational", "slug": "rational", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/88c9150f7633337a9e66431f2838539b473b583f.png", "website": "https://rational.to", "all_locations": "San Francisco, CA, USA", "long_description": "Rational is building the first accounting firm with no humans. It's entirely run by our AI employees, who go from client intake to doing the books to filing with the IRS. We secure 90% cost reduction, 24/7 support, and an overall seamless experience for businesses.", "one_liner": "First Zero-human Accounting Firm", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Operations", "launched_at": 1784172515, "tags": [ "Finance", "B2B", "Enterprise", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Operations" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/rational", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/rational.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/rationalto", "x_url": "https://x.com/RationalDotTo", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/88c9150f7633337a9e66431f2838539b473b583f.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1591705, "full_name": "Jibril Moinuddin", "title": "Founder/CEO", "founder_bio": "CEO at Rational. Previously founding engineer at Buckeye AI, TrueFoundry. Founded STEMplicity Education. Dropped out of University of Michigan. Non-Trivial Fellow.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jibril-moinuddin-32702b271/", "x_url": "https://x.com/JibrilAsif", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9218a0f71bb3d439dfb97ee0da080509800aefe7.jpg" }, { "user_id": 3411828, "full_name": "Christ Xu", "title": "Founder/CTO", "founder_bio": "CTO at Rational. Prev Youngest Tencent Senior SWE. 2x IOI Camp, IMO Camp, published algo researcher.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/christ-xu-687868298", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/23ba5c0e4852292e86dfe9667dd895b8c678ad64.jpg" } ], "jobs": [ { "id": 106940, "title": "Forward-Deployed Engineer", "url": "https://www.ycombinator.com/companies/rational/jobs/FGBkhzz-forward-deployed-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$120K - $200K", "equity_range": "0.50% - 2.00%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 106941, "title": "Machine Learning Researcher ", "url": "https://www.ycombinator.com/companies/rational/jobs/vRTNbJi-machine-learning-researcher", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Machine learning", "salary_range": "$150K - $200K", "equity_range": "0.50% - 2.00%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 106942, "title": "Software Engineering Intern", "url": "https://www.ycombinator.com/companies/rational/jobs/6ycNzJc-software-engineering-intern", "location": "San Francisco, CA, US", "type": "Internship", "role": "Engineering", "role_type": "Full stack", "salary_range": "$8K - $12K / monthly", "equity_range": "", "experience": null, "visa": "US citizen/visa only", "skills": [] }, { "id": 106944, "title": "Applied AI", "url": "https://www.ycombinator.com/companies/rational/jobs/u1XiPpU-applied-ai", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Machine learning", "salary_range": "$120K - $200K", "equity_range": "0.50% - 2.00%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] } ], "news": [ { "title": "Rational Launches AI Employees for Accounting Firms", "url": "https://insidepublicaccounting.com/2026/07/17/rational-launches-ai-employees-for-accounting-firms/", "date": "Jul 17, 2026" } ], "launches": [ { "id": 114437, "title": "Rational - First Zero-human Accounting Firm", "tagline": "Agents will be trusted.", "body": "\n\n**We’re disrupting the disruptors.**\n\nYou’ve heard of so many AI accounting startups already. These new “AI-native” accounting firms are still designed to be run by humans in two facets: the first 25%, fostering client relationships, and the final 25%, signing off and filing with the IRS.\n\n**Josh Kushner’s company, Thrive Holdings, the largest disruptor in this space, explicitly argues that agents will never replace this** \\[1\\]**.** Their reasoning? Agents don’t carry trust.\n\nCurrently, agents are not trusted, humans are. That’s why every AI startup nowadays does the work, and leaves judgment to a human. But this is changing, and we’re early. As new agent-centric companies grow (Instinct, Natural, AgentMail), agents become common in daily life and work, and eventually will become trusted entities \\[2\\].\n\nOnce agents are trusted, there is no reason they cannot own entire, traditionally trust-heavy industries, end-to-end. Rational is building for that world.\n\n**Reach out to [founders@rational.to](mailto:founders@rational.to) or visit [rational.to](http://rational.to) if interested in learning more.**\n\n---\n\n\\[1\\] . A fraction of the market does care about trust and the human relationship with their accountant, but most of us definitely do not want to spend a second more with one than we have to.\n\n\\[2\\] These companies inherently bet on agents to proliferate in society. We also believe AI will continue to get better and safer. Since trust is achieved through consistent, positive actions, and it is in the interest of the entire tech industry for their agents to produce consistent positive actions, our firm will be the first to enjoy the principles to do concrete work in a massive market.", "url": "https://www.ycombinator.com/launches/Tll-rational-first-zero-human-accounting-firm", "votes": 22, "created_at": "2026-09-09T19:03:11.387Z" }, { "id": 112573, "title": "Rational - AI Employees for Accounting Teams", "tagline": "AI employees for in-house accounting teams", "body": "Rational builds AI employees for in-house accounting teams. They onboard themselves, handle real work within existing tools, and keep a human in control of every decision that matters. \n\nOur AI employees conduct interviews with each role on your team, creating a mental map of your workflows, inefficiencies, and who to route approvals to. After you connect them to your current stack, they simply do the work, 24/7. They learn from every single interaction and become more accurate over time — and their insights are completely isolated for your team. They’re truly your AI employees.\n\nWe’re live with major enterprise customers, including a Fortune Global 100.\n\n\n\nToday, we’re also releasing our Web Playground for early access.\n\n**Visit rational.to/book-demo to get access.**\n\nNot part of an accounting team but know someone in PE, advising, or accounting? We would love to talk. Email [founders@rational.to](mailto:founders@rational.to).", "url": "https://www.ycombinator.com/launches/THh-rational-ai-employees-for-accounting-teams", "votes": 3, "created_at": "2026-08-28T16:00:00.566Z" }, { "id": 106060, "title": "Rational: AI Employees for Accounting", "tagline": "24/7 hires to solve the talent shortage within your firm's existing stack", "body": "Rational builds AI employees for accounting and tax firms. They onboard themselves to your firm, handle real work within your existing tools, and keep a human in control of every decision that matters — so firms can finally say yes to the clients they've been turning away.\n\n\n\n_The problem_\n\nAccounting has an economics problem: demand is there, capacity isn't. Most firms turned away paying clients last year simply because they didn't have the people to do the work, and the talent pool keeps shrinking as accountants leave the profession faster than firms can replace them. More software hasn't fixed this — and a chatbot that summarizes a spreadsheet definitely hasn't. The parts of accounting that actually matter, like judgment, trust, and sign-off, should stay human. Everything around them shouldn't have to be a bottleneck.\n\n_Our solution_\n\nRational’s AI employees conduct interviews with each role on your team, creating a mental map of your firm’s workflows, inefficiencies, and who to route approvals to. After you connect them to your current stack, they simply do the work, 24/7. They learn from every single interaction and become more accurate over time — and their insights are completely isolated for your firm. They’re truly _your_ AI employees.\n\nRational already works with every firm size: from individuals to Fortune Global 500 companies. Doing everything from document-chasing to research to tax preparation, our employees supplement your firm’s labor without the significant onboarding curve or cost of a new hire.\n\n_Book a Demo_\n\nStop saying no to clients or retaining inefficiencies in your firm. Hire your first AI employees today, at .\n\nNot an accounting firm but know someone in PE, advising, or accounting? We would love an introduction. Email founders@rational.to.", "url": "https://www.ycombinator.com/launches/Rae-rational-ai-employees-for-accounting", "votes": 18, "created_at": "2026-07-17T17:45:59.589Z" } ] }, { "id": 32623, "name": "Omanta", "slug": "omanta", "former_names": [ "ClinSIGHT AI" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ae1812a6e6a806880355a25e542d3970bd58deb5.png", "website": "https://omanta.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Omanta is a personalized research lab for patients who have exhausted standard options, want to prepare proactively for recurrence, or are living with an unresolved diagnosis.\r\n\r\nWe direct the full operational infrastructure, personnel, expertise, and technology of a large research program toward understanding and addressing the challenges of individual patients across oncology, rare disease, and undiagnosed conditions.\r\n\r\nFor clinicians, we act as a collaborative partner, providing a living roadmap for ongoing patient care built on a rigorous and evolving scientific evidence base.\r\n\r\nWe fully operationalize patient-specific research programs, covering tissue logistics, diagnostic follow-up, data integration, multiomic analysis, therapeutic strategy development, and more.\r\n\r\nOmanta brings context and clarity to complex cases, helping clinicians see more of the scientific picture as they decide how best to care for their patients.", "one_liner": "A research lab for one", "team_size": 6, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1785793252, "tags": [ "Digital Health", "Healthcare", "Health & Wellness", "Oncology", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/omanta", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/omanta.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/omanta/", "x_url": "https://x.com/Omantahealth", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ae1812a6e6a806880355a25e542d3970bd58deb5.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3670159, "full_name": "Alfredo Gonzalez", "title": "Founder/CEO", "founder_bio": "Co-Founder and CEO of Omanta. PhD in Cancer Bioinformatics from UCLA, biomedical researcher at Broad Institute and Harvard. Spearheaded personalized therapeutic campaigns for patients creating personalized vaccines, immunotherapies, and radioligands. 10+ years in personalized medicine. Fueled by the belief that the future of Health is deeply personalized.\n\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/alfredo-gonzalez-phd/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9d5790791f1a19912cd701c3817cb00d66003eca.jpg" }, { "user_id": 3399110, "full_name": "Ranad Humeidi", "title": "Founder and Chief Scientific Officer", "founder_bio": "Co-founder & CSO at Omanta, working with patients to turn frontier science into personalized treatment strategies. Harvard Chemical Biology PhD in 3 years; identified a druggable ulcerative colitis mechanism that led to a multimillion-dollar GSK collaboration. Built and directed a multimillion-dollar rare cancer program for one patient, coordinating researchers, physicians, biotechs, and drug developers. Led strategies for 12+ complex patients after standard options were exhausted.\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rhumeidi", "x_url": "https://x.com/Ranadhumeidi", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4e2478bf2c078b78d7428a780e36bd00d7bba74a.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108925, "title": "Omanta: A personal research lab dedicated to each patient.", "tagline": "A research lab, for one.", "body": "Hey YC! We're Alfredo and Ranad, the founders of Omanta.\n\n**TL;DR: Omanta creates a personalized research lab around a single patient. We integrate a patient’s complete medical record, personal genomics, and use AI to take on the most complex cancer cases. Our researchers tackled GitLab co-founder Sid Sijbrandij’s case, helping bring him into complete remission.**\n\n\n\n\\\n**Why this matters**\\\nIf you're a cancer patient, you'd think the hardest cases have entire teams dedicated to solving them. The reality: most doctors have about 20 minutes per patient, while the science never stops moving. New discoveries emerge every day, yet reaching standard clinical care can take over a decade (\\~17 years on average). Physicians make high-stakes decisions with limited time, and even the most difficult cases rarely have anyone whose full-time job is to interrogate every layer of that one patient's disease and push the newest science toward their treatment.\n\nThat's the gap we built Omanta to close.\\\n\\\n**What we do**\\\nWe meet patients wherever they are in their journey. We integrate their complete medical record and personal genomics with AI and clinical evidence to understand the molecular drivers of their disease. Then we map every existing therapy against their biology, and if the right therapy doesn't exist, we build it, from cancer vaccines and immunotherapies to highly targeted radioligand therapies.\\\n\\\nWe've worked on some of the most complex cancer cases in the world. Most notably, **Alfredo was the genomics lead for the personalized therapeutic campaign behind GitLab founder Sid Sijbrandij's complete remission**, partnering with leading institutions to pursue personalized mRNA/DNA vaccines, CAR-T programs, drug repurposing, expanded-access therapies, and continuous molecular monitoring. We've seen what becomes possible when the full force of modern science is organized around a single person, and Omanta brings that depth of investigation to every case.\\\n\\\n**Why us**\\\nWe met eight years ago working on early CRISPR cancer research and have been building toward this ever since.\n\n* **Alfredo Gonzalez, PhD (CEO)** — translational genomics. Led the year-long investigation of 25TB of sequencing and molecular data for Sid Sijbrandij, key missing information in existing genomic reports, and uncovered actionable findings that reshaped his therapeutic strategy, including personalized mRNA/DNA vaccines, CAR-T, expanded access, drug repurposing, and continuous monitoring. PhD at UCLA, where he built one of the world's largest centralized multi-omic datasets for rare tumors. Previously at the Broad Institute (Cancer Dependency Map, validating targets from large-scale CRISPR screens) and Harvard's Wagers Lab.\n* **Ranad Humeidi, PhD (CSO)** — chemical biologist. Earned her PhD in Chemical Biology from Harvard in three years, where she identified a druggable mechanism in ulcerative colitis that became the basis of a multimillion-dollar GSK collaboration. Built and directed a multimillion-dollar scientific program centered on a single rare-cancer patient, then went on to direct scientific strategy for 12+ medically complex patients after standard options were exhausted. At the Broad, she co-created and co-led a fellowship turning scientific discoveries into companies.\n\n\\\n**Our ask**\n\n* Intros. If you or someone you love is facing a difficult cancer diagnosis, we want to hear from you.\n* We're also looking to connect with physicians who want a dedicated research team behind their most complex cases.\n* Reach us here on Bookface or at [alfredo@omanta.com](mailto:alfredo@omanta.com) & [ranad@omanta.com](mailto:ranad@omanta.com).\n\n[omanta.com](http://omanta.com)\\\nA research lab, for one.", "url": "https://www.ycombinator.com/launches/SKr-omanta-a-personal-research-lab-dedicated-to-each-patient", "votes": 73, "created_at": "2026-08-04T13:31:15.555Z" } ] }, { "id": 32624, "name": "Fabraix", "slug": "fabraix", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/d3d83aa2013bbe6b22b37f316b294d0c717eea35.png", "website": "https://fabraix.com", "all_locations": "San Francisco, CA, USA", "long_description": "Fabraix builds state-of-the-art AI red-teaming agents that continuously detect security vulnerabilities in customer-facing AI. Our product, Nyx, has already found vulnerabilities in agents at dozens of Fortune 500 companies. On AgentHarm, the leading benchmark for offensive AI security, Nyx achieved a 78% attack success rate, compared with 67% for GPT-5.6 Sol.\r\n\r\nAI agents can change more often than teams can test them. A new model, prompt, tool, permission, or data source can change what an agent does, even when the application code stays the same. And AI is also increasing how much software companies produce and how often it changes. We built Nyx to automate the work required to red-team AI agents. It is multi modal by design and interacts with the target indirectly through controlled replicas of SaaS products and websites that we maintain, placing malicious payloads in webpages, documents, files, messages, and tool outputs to test how an agent behaves under different conditions.\r\n\r\nNyx often finds its first vulnerability within minutes or hours rather than days or weeks. Because the work is fully automated, companies can repeat the test with every change and at a much lower cost than what a 7-figure red-teaming (and point-in-time) engagement would require.\r\n\r\nWe're also the team behind ACE (Adversarial Cost to Exploit), a benchmark that measures AI security in terms of how much it costs attackers to break an AI system; providing a game-theoretic framework to understand how motivated a rational attacker would be in exploiting the system.", "one_liner": "The world's frontier hacker of AI agents.", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Security", "launched_at": 1783913331, "tags": [ "Reinforcement Learning", "Cybersecurity", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Security" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/fabraix", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/fabraix.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/fabraix/", "x_url": "https://x.com/FabraixHQ", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/d3d83aa2013bbe6b22b37f316b294d0c717eea35.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 693526, "full_name": "Ahmed Aly", "title": "Founder/CEO", "founder_bio": "Co-founder @ Fabraix, building a frontier hacker agent that finds failure modes and security exploits in AI systems before real users do. Previously the first data scientist at Sequoia-backed Two, where he built AI systems for adversarial environments from scratch and scaled them to process $1B+ in annual B2B transactions, averting $50M in losses for multiple Fortune 150 companies. A published researcher with multiple papers in Q1 journals and a UCL PhD drop-out.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ahmed-a-657264b4/", "x_url": "https://x.com/zachdotai", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/72bc6e2cde8727671fbf9c026c0c9a0ffcc80f4f.jpg" }, { "user_id": 2769430, "full_name": "Ibrahim Abdu", "title": "Founder/CTO", "founder_bio": "Red-teamed AI agents in Meta's Privacy org; built an SRE agent triaging 3M+ errors/month, saving 10,000+ engineering hours a year. Built a compiler for insider-trading detection at TradingHub: $100M+ in daily trades; Oxford PPE, top 8% of cohort.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/i-a-7352b7137/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e3d923fd8bbb4ccfb176340c73f6ebc968516321.jpg" } ], "jobs": [], "news": [ { "title": "Meet The YC Startups Betting On What Comes Next", "url": "https://www.forbes.com/sites/dariashunina/2026/09/03/meet-the-yc-startups-betting-on-what-comes-next/", "date": "Sep 03, 2026" } ], "launches": [ { "id": 106270, "title": "Fabraix: The Frontier Agent that Hacks Customer-facing AI.", "tagline": "Fabraix builds AI red-teaming agents that continuously detect security vulnerabilities in customer-facing AI. We already found vulnerabilities in agents at dozens of Fortune 500 companies.", "body": "**TL;DR**: Fabraix builds AI red-teaming agents that continuously detect security vulnerabilities in customer-facing AI. Our product, Nyx, has already found vulnerabilities in agents at dozens of Fortune 500 companies, and hits a 78% attack success rate on AgentHarm (leading offensive AI security benchmark) vs. 67% for GPT-5.6 Sol.\n\n**Ask**: If your team ships customer-facing AI agents (chat, voice, browser, or coding) and needs to continuously test them for security vulnerabilities, find us at [founders@fabraix.com](mailto:founders@fabraix.com).\n\n\n\n**The Problem**\n\nAI agents introduce new vulnerabilities not present in traditional software..\n\nIn order for agents to be truly useful, they must be trusted to interact with untrusted surfaces like MCPs, websites and files, take real actions, and have access to confidential data. You can’t tell your defenses are sound by just reading your prompts since the agent's behaviour is not deterministic. The best thing to do is run your agent and really try to break it.  \n\nDoing this red-teaming process manually is time consuming, costly, and you are unlikely to catch everything. And as soon as the agent is updated, you need to repeat it. Security teams simply don’t have the bandwidth to keep up with the rapid deployment inside organisations.\n\n**Our Solution**\n\nNyx is our red-teaming agent that tests customer-facing AI agents, via direct interaction and also indirectly via its environment. It requires no access to your agent’s source code in order to test.\n\nNyx is highly adaptive when attacking. Rather than cycling through hard coded payloads like existing tools, it dynamically adjusts its attack strategies based on your agents defenses and responses. It probes your agent over a wide surface area of attack vectors, and deep dives This adaptive approach makes it 20x more effective as a red-teamer compared to other solutions.\n\nWe maintain a library of over 10,000 attack strategies and jailbreaks. These are high level strategies which are used by Nyx to generate dynamic adaptive attacks, over multiple turns. \n\nEvery finding includes the attack steps, the agent’s responses, and the resulting failure for you to reproduce. You can then rerun the same test after a release to check whether the failure remains fixed. \n\nOn the AgentHarm benchmark, Nyx achieved a 78% attack success rate, compared with 67% for GPT-5.6 Sol. Nyx has also found exploitable failures in public-facing agents operated by dozens of Fortune 500 companies.\n\nWe also built ACE (Adversarial Cost to Exploit), a benchmark that measures AI security in terms of how much it costs attackers to break an AI system; providing a game-theoretic framework to understand how motivated a rational attacker would be in exploiting the system.\n\n**The Team**\n\nWe both have a lot of experience building and securing AI agents at scale.\n\nIbrahim built AI agents at Meta that diagnosed and fixed production errors. Before that, he built compilers and database engines in Fintech. He graduated from Oxford in the top 8% of his cohort.\n\nAhmed led international payments fraud team at Monzo. Before that, he was the first data scientist at a Sequoia-backed startup, where he built its fraud engine from scratch into processing more than $1B in annual B2B transactions and preventing tens of millions in losses for Fortune 150 companies. He has published several peer-reviewed research papers in Q1 journals.\n\nWe previously worked together, and have known each other for almost four years.\n\n![uploaded image](/media/?type=post&id=106270&key=user_uploads/693526/5f437cbd-2009-40cc-8c69-746da9ea8300)\n\n**The Ask**\n\nIf you ship a customer-facing (chat, voice, browser, or coding) agent, we can help you find all the failure modes and built a self-healing loop to discover, triage, and maintain secure AI agents.\n\n[founders@fabraix.com](mailto:founders@fabraix.com)", "url": "https://www.ycombinator.com/launches/Re2-fabraix-the-frontier-agent-that-hacks-customer-facing-ai", "votes": 127, "created_at": "2026-07-20T15:06:58.917Z" } ] }, { "id": 32625, "name": "Qlo", "slug": "qlo", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/6bdf97a8fbe4a8a51a745ce366d9c6c2bfc6d75c.png", "website": "https://www.getqlo.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Qlo builds AI agents to run commercial insurance underwriting. We start by fully automating the operational work around risk decisions, become the system of record, and earn autonomy workflow by workflow toward the autonomous AI underwriter.\r\n\r\nWe’re rebuilding underwriting infrastructure from the ground up so human underwriters and AI agents can share the same accounts, rules, documents, permissions, and workflow state. \r\n\r\nWe believe every commercial insurance program will eventually be built and operated end to end by AI agents on Qlo.", "one_liner": "Legora for commercial insurance carriers", "team_size": 2, "industry": "Fintech", "subindustry": "Fintech -> Insurance", "launched_at": 1786080312, "tags": [ "Insurance", "Enterprise Software", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech", "Insurance" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/qlo", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/qlo.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/getqlo/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/6bdf97a8fbe4a8a51a745ce366d9c6c2bfc6d75c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 1640725, "full_name": "Ryan Hu", "title": "Founder/CEO", "founder_bio": "Founder of Qlo. Previously bootstrapped a B2B product studio to $500K ARR and built a commercial insurance marketplace that received an acquisition offer from a public insurer. Moved from Taiwan to build in the U.S. Studied Electrical Engineering at National Taiwan University before dropping out.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/chenshuohu/", "x_url": "https://x.com/hcsfox", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/088d399ab10b6ca1a2786d60505cd551af8dd69a.jpg" }, { "user_id": 3743083, "full_name": "Liv Huang", "title": "Founder", "founder_bio": "Co-Founder at Qlo. Previously Cloud Engineer at CuboAi, building scalable backend infrastructure and driving cross-functional AI adoption. Self-taught engineer with an academic background in Counseling and Educational Psychology.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/livhg", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/78b8b2063eb5fea134f67a8cc93fc8d3c4f45897.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111422, "title": "Qlo - AI agents that run commercial underwriting", "tagline": "The underwriting system of record for the AI era", "body": "**TL;DR:** Qlo uses AI agents to run commercial underwriting operations from submission to renewal. We rebuilt the underlying system so agents can operate across the entire workflow, not sit beside another legacy system. One enterprise client signed a two-year contract, went live in May, and has since tripled weekly primary quote throughput. The end state is autonomous underwriting that follows each company’s underwriting judgment and stays within the carrier’s appetite.\n\nToday, we’re launching the **Qlo Inbox Agent**.\n\n\n\n### The Problem\n\nMore than $110B in commercial insurance premium flows through U.S. MGAs every year. Much of the work still begins with an email from a broker.\n\nCommercial underwriters can spend up to 80% of their day in the inbox, chasing missing applications, loss runs, and other submission documents. Even a submission with roughly $50K in annual premium can take 48 hours to turn into a quote.\n\nOnce the information arrives, teams still need to prepare the data, run the rater, and assemble the right endorsements and quote variants.\n\nWhen volume grows, teams add headcount. But underwriting judgment is hard to train and keep consistent. Costs rise, good submissions remain untouched, and inconsistent decisions can hurt loss ratios.\n\nMore premium left on the table, now with more people on payroll.\n\n### What We Built\n\nQlo runs the commercial underwriting workflow from submission to renewal. Specialized agents handle each operational step, while supervisor agents coordinate the work and communicate with the underwriter.\n\nThe Inbox Agent is one of those supervisors. An underwriter can message it from a phone like a teammate. It understands the instruction, calls the right subagents, and completes the work inside Qlo.\n\n**The underwriter makes the call. The agents make it happen.**\n\nOne of those subagents is the Rater Converter we launched on Bookface. It turns an existing rating spreadsheet into a validated digital worksheet the rest of the workflow can use.\n\nThis is what agent-compatible means to us: agents understand the workflow, have the right context, and can act across the system. Today, they execute the underwriter’s decisions. The end state is autonomous underwriting within each company’s judgment, appetite, and controls.\n\n**The party has begun. Let the agents dance on top of Qlo.**\n\nGuidewire and Vertafore helped define the last era of commercial insurance software. We’re building Qlo to define the next one.\n\n### Why Us\n\nBefore Qlo, Ryan spent 2.5 years building a commercial insurance marketplace and saw carrier-side systems up close. The company received acquisition terms from a public insurance company, and two founders who were customers there are now advisors to Qlo.\n\nLiv previously led backend at a Taiwanese baby-monitor company generating roughly $15M in annual revenue. Ryan and Liv had already built together before starting Qlo.\n\nWe build alongside experienced underwriters who want to help define how commercial insurance will operate in the AI era.\n\n### Our Ask\n\nWe’re looking for:\n\n* **Introductions to commercial MGAs and program administrators**, especially submission-heavy programs with complex rating files or underwriting teams hitting the limits of their current systems.\n* **Introductions to investors who understand commercial insurance** and see the same path toward autonomous underwriting.\n* **Insurance operators with strong opinions about the future of underwriting**. Even if you’re not looking for a new system, we’d still like to learn how your business works.\n\nIf you run an MGA or program, want to build with us, or see the same future as an investor, email us at [founders@getqlo.com](mailto:founders@getqlo.com).\n\nCommercial insurance moves slowly because **distribution beats architecture**.\n\nBut distribution eventually stops scaling when the architecture underneath it no longer works.\n\nThe system that solves the real problem will be adopted. Time will tell.\n\nWe’ll keep building.\n\n![uploaded image](/media/?type=post&id=111422&key=user_uploads/1640725/d836b808-6597-42f6-b85c-a1aed3413267)\n\n", "url": "https://www.ycombinator.com/launches/Sz8-qlo-ai-agents-that-run-commercial-underwriting", "votes": 35, "created_at": "2026-08-20T15:00:00.451Z" } ] }, { "id": 32627, "name": "Tenor", "slug": "tenor", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/28b2676e2b5c4bb6bce97d4ce827f594e9743ce2.png", "website": "https://heytenor.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Tenor is a full-stack infrastructure for building and running AI-native businesses. We’re not just an agent framework, we manage the entire loop from how AI executes work to how every action, decision, and business outcome is captured and learned from. Owning the full stack lets companies deploy governed AI workflows, understand what actually drives results, and continuously improve how their business operates without building the infrastructure themselves.", "one_liner": "Infrastructure for building and scaling AI-native businesses.", "team_size": 3, "industry": "B2B", "subindustry": "B2B", "launched_at": 1786480856, "tags": [ "B2B", "Automation", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/tenor", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/tenor.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/heytenor/", "x_url": "https://x.com/agentemployees", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/28b2676e2b5c4bb6bce97d4ce827f594e9743ce2.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 536910, "full_name": "Hamze Al-Zamkan", "title": "Founder/CEO", "founder_bio": "Co-Founder @ Tenor", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/hamze-al-zamkan/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f94e739a9abd5fd7d16c859836f4d5327eda19ca.jpg" }, { "user_id": 2719109, "full_name": "Amgad Al-Zamkan", "title": "Founder/CPO", "founder_bio": "Co-founder @ Tenor.\nPreviously at Amazon, studied CS at UC Berkeley and TUM. As a grad student, Amgad's research focused on recommendation algorithms and shaping adaptive decision systems.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/amgad-al-zamkan/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f05cd43d0888166e77a571f9543b399a24658f4c.jpg" }, { "user_id": 989310, "full_name": "Muhtasham Oblokulov", "title": "Founder/CTO", "founder_bio": "Co-founder @ Tenor. \nEarly contributor to DeepSeek and co-author of the highly cited StarCoder series and CodeClash with Stanford University. Previously deployed large-scale AI systems across global reinsurance markets.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/muhtasham/", "x_url": "https://x.com/Muhtasham9", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a132f91a426d3f455e5c367a7d8125d02749c418.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110150, "title": "Tenor: AI that owns jobs end-to-end and is accountable for business outcomes.", "tagline": "Deploy AI that owns complete jobs and are measured against predefined outcomes and KPIs. Tenor's AI operates across your tools and improves its decisions with every outcome to drive better results. ", "body": "**TL;DR:**\n\nThe defining resource of the internet was page views. The defining resource of AI is tokens. But while companies can measure exactly what their tokens cost, they still can’t easily tell what those tokens produced. \\\nTenor connects the two. Every piece of AI work starts with an outcome and a budget. Tenor tracks the resources spent against the result produced, learns from that result, and helps determine where the next tokens should go.\\\nThat’s the bigger opportunity: not just attributing AI spend, but allocating machine intelligence toward what creates value. Once a company knows where its tokens make money, it can do more of what works, less of what doesn’t, and continuously reshape itself around where AI creates value.\n\n## 🤯 The Problem\n\nWhile companies have become good at measuring AI consumption, they still can’t tell what that intelligence produced in business value.\\\n\\\nThe internet went through a similar transition.\n\nIts defining resource was page views. Early on, advertisers could buy impressions and measure exposure. Search advertising changed the model by connecting spend to intent and outcomes. Google then turned that into an engine for allocating dollars toward what performed.\n\nAI is still early in that transition. Companies deploy copilots, agents, and automations across the organization. Each produces activity and consumes tokens. But the economic context usually lives somewhere else.\n\nWhat was this intelligence supposed to accomplish? What changed because it ran? Was the outcome worth what it cost? Should the next token go here or somewhere else?\n\nWithout those answers, AI remains largely a software expense: something companies buy and monitor rather than a resource they can economically allocate.\n\nAnd the bigger loss isn’t wasted tokens. It’s lost learning.\n\nMillions of pieces of work happen, outcomes disappear into different systems, and very little of that experience determines how the next unit of intelligence gets deployed.\n\n### ⚙️ The Solution\n\nTenor starts with the outcome.\n\nBefore intelligence is deployed, the work gets an economic identity: what it owns, what success means, what it can spend, which systems it can act on, and when humans need to intervene.\n\nNow every token is spent in service of something measurable.\n\nSo instead of: This agent consumed X.\n\nYou can ask: What did X produce?\n\nThat closes the loop: allocate → work → outcome → attribute → learn → reallocate\n\nAttribution tells you what worked. Allocation is what makes that knowledge valuable.\n\nWork that produces value can receive more intelligence. Work that doesn’t can be changed, rebuilt, or stopped. Every outcome improves the next allocation decision.\n\nOver time, that feedback loop reaches beyond individual workflows.\n\nResponsibilities move. Processes disappear. New ones emerge. Software changes around the work. The organization gradually restructures itself around where machine intelligence produces the greatest value.\n\nThat’s what we’re building Tenor for: not just to measure the token economy, but to give companies the machinery to operate inside it.\n\n## 👋 Team\n\nMuhtasham Oblokulov — AI researcher, 3,000+ citations, ICML-published. Former Senior ML Engineer at Munich Re, where his models drove $1B in revenue.\n\nHamze Al-Zamkan — President of [TUM.ai](http://TUM.ai), which he grew into Europe's largest student-led AI organization and a talent pipeline for Munich's AI ecosystem. Deployed AI systems at appliedAI, and worked on growth at Voggt and Terra. Studied CS at TUM and CDTM.\n\nAmgad Al-Zamkan — Studied CS at TUM and UC Berkeley. Researched adaptive decision systems. Automated judgment-heavy work at Amazon.\n\n![uploaded image](/media/?type=post&id=110150&key=user_uploads/2719109/0a5e6384-5ed4-4345-9e88-5c3114a5ebe0)\n\n## 🙏 Our Ask\n\nIf your AI spend is growing and your company isn't visibly getting better at its work every week, that gap is our market.\n\nWant to see what the work produces? email [hamze@heytenor.com](mailto:hamze@heytenor.com)", "url": "https://www.ycombinator.com/launches/Sec-tenor-ai-that-owns-jobs-end-to-end-and-is-accountable-for-business-outcomes", "votes": 50, "created_at": "2026-08-12T00:00:00.512Z" } ] }, { "id": 32649, "name": "Daqstra", "slug": "daqstra", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/dad9cd6f305ec9c8a219f22ee7d83e68f227d245.png", "website": "https://www.daqstra.com", "all_locations": "San Francisco, CA, USA", "long_description": "Daqstra is the AI-native orchestration layer for physical R&D. It saves engineers in Aerospace, Energy and Robotics hundreds of man hours managing test infrastructure by connecting fragmented test hardware, bringing live and historical test data into one place, and by giving engineers an API to orchestrate their entire test setup. We enable teams to spend less time rebuilding test infrastructure and more time improving the hardware they test.", "one_liner": "AI-native test infrastructure layer for hardware companies", "team_size": 3, "industry": "B2B", "subindustry": "B2B", "launched_at": 1784276991, "tags": [ "Hard Tech", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/daqstra", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/daqstra.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/daqstra/", "x_url": "https://x.com/daqstra", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/dad9cd6f305ec9c8a219f22ee7d83e68f227d245.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 2554874, "full_name": "Darell Chua", "title": "Founder/CEO", "founder_bio": "Founder at Daqstra (YC S26). Previously Systems Performance Architecture at Rivian, Member of Technical Staff at DSO National Labs Singapore. MS in Computer Engineering from UC San Diego. Developed end-to-end high performance hardware and software systems for aerospace and drone applications.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/darell-chua/", "x_url": "https://x.com/heyitsdarell", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fcda896cbacc71f741df36a68755f8fdc284f41f.jpg" }, { "user_id": 2547322, "full_name": "Anchit Kumar", "title": "Founder/CTO", "founder_bio": "I’m a co-founder of Daqstra (YC S26), building AI-native orchestration for test infrastructure in physical R&D. I studied computer science at UCSD, where I led AI and autonomy for Mars cave exploration drones and worked on multimodal AI safety for vision-language-action robots. My co-founder and I met building rocket engines, and I’ve since worked across 10+ test campaigns in the Mojave Desert.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/anchitk/", "x_url": "https://x.com/anchit__kumar", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a52b0a3b2d8e6caf3372ee158748700e91fa9ce5.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 106987, "title": "Daqstra – Test infrastructure shouldn't bottleneck great engineering", "tagline": "AI-native orchestration for physical R&D testing", "body": "# \\*\\*\\*\\*\n\n**TL;DR: [Daqstra](https://www.daqstra.com/) is the AI-native test orchestration platform for physical R&D. We connect fragmented test hardware, capture every run, and give engineers an API to operate their test setup, helping aerospace, energy, and robotics teams spend less time rebuilding infrastructure.**\n\nHey! We’re [Anchit Kumar](https://www.linkedin.com/in/anchitk/) and [Darell Chua](https://www.linkedin.com/in/darell-chua/), the founders of [Daqstra](https://www.daqstra.com/).\n\n![uploaded image](/media/?type=post&id=106987&key=user_uploads/2547322/7ed7475e-d673-4796-9462-d40f72da12ad)\n\nWe met while testing liquid rocket engines in the Mojave Desert.\n\nThe engines were difficult, but much of the friction came from the infrastructure around them. Before each test, we connected sensors, custom electronics, NI hardware, PLCs, and controllers that were never designed to work together. After each run, the data was scattered across machines, file formats, scripts, and people’s heads.\n\nBefore we could learn from one test, we often had to reconstruct what had happened.\n\nAfter talking with teams building rockets, energy systems, robots, and drones, we kept hearing the same story: **their hardware was advancing faster than the infrastructure used to test it.**\n\n## What We Built\n\n[Watch the demo](https://www.youtube.com/watch?v=ykvD67omalI) or visit [daqstra.com](http://daqstra.com).\n\nDaqstra works with the equipment teams already have, including NI/LabVIEW, PLCs, custom DAQ systems, sensors, controllers, and vendor hardware.\n\n* **Daqstra Edge** connects hardware at the test site and records commands alongside the resulting telemetry.\n* **Daqstra Hub** stores each run with its configuration, events, notes, and engineering context, making it easier to monitor tests, compare runs, and investigate anomalies.\n* **Daqstra API** lets engineers configure hardware and execute repeatable test sequences through software.\n\nTogether, these give teams one interface across otherwise incompatible hardware without forcing them to replace their existing stack.\n\n## What We’re Already Doing\n\nWe’re working with advanced hardware teams across aerospace, energy, and drones. Our active aerospace programs use Daqstra day to day across fragmented, multi-vendor test environments.\n\nIn energy, we’ve signed a pilot with a palm oil processor targeting $100K+ in annual savings by reducing recoverable oil lost through its effluent stream. We’re also helping drone teams unify telemetry, control, and test data across vehicles and ground systems.\n\n## Who We’d Love to Meet\n\nIf you lead test, propulsion, integration, validation, or hardware engineering and run frequent instrumented tests, we’d love to talk.\n\nDaqstra is especially useful if your team:\n\n* Works across vendor equipment, PLCs, custom electronics, and in-house software\n* Reconfigures its setup between tests or struggles to compare runs\n* Relies on a few engineers who understand the entire test environment\n\nWe’re initially focused on aerospace, energy, drones, and robotics. If this sounds like your team, reply here or email [**team@daqstra.com**](mailto:team@daqstra.com).\n\nIf it sounds like someone you know, we’d appreciate an introduction to the person responsible for test infrastructure, integration, or validation.\n\n**Ad Astra Per Daqstra**", "url": "https://www.ycombinator.com/launches/Rpb-daqstra-test-infrastructure-shouldn-t-bottleneck-great-engineering", "votes": 10, "created_at": "2026-07-23T19:42:19.138Z" } ] }, { "id": 32650, "name": "Risklytics", "slug": "risklytics", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/879622d4fa103941586542ac659eaf9b60d897f0.png", "website": "https://www.risklytics.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Risklytics is the insurance company for frontier tech such as physical AI, robotics, and data centers. Companies answer 6 questions and we bind a tailored policy that other insurance companies won’t.", "one_liner": "AI-Native Insurance for Frontier Tech", "team_size": 2, "industry": "Fintech", "subindustry": "Fintech -> Insurance", "launched_at": 1783123197, "tags": [ "Hardware", "Machine Learning", "B2B", "Insurance", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech", "Insurance" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/risklytics", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/risklytics.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/risklytics", "x_url": "https://x.com/Risklytics", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/879622d4fa103941586542ac659eaf9b60d897f0.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 2973325, "full_name": "Samuel Gold", "title": "Founder/CEO", "founder_bio": "Founder at Risklytics (YC S26). Building a brokerage for frontier tech. On leave from Harvard.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/samuelggold", "x_url": "https://x.com/SamGGld", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b99591b78044a8c663b7b4eb65a1fe000a6bb179.jpg" }, { "user_id": 2666435, "full_name": "Alexander Risio", "title": "Founder/CTO", "founder_bio": "Founder at Risklytics (YC S26). Building a new type of insurance brokerage.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/alexanderrisio/", "x_url": "https://x.com/alexrisio", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/dc585a4120c666a66e51d372101b7f5ed304f437.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110197, "title": "Risklytics: Insurance for the Frontier", "tagline": "Insurance for companies pushing the frontier of what is possible with AI, from robot fleets to agents in production.", "body": "**TLDR:** Risklytics is a commercial insurance brokerage for companies building with AI in the physical space, from robot fleets on customer sites to agents running in production. This year many insurance companies stopped covering risks that involve AI. We find the insurers that will still cover it, translate what you're building into terms they can price, and read every clause before anything is signed.\n\n\n\n**The problem**\n\nStarting this year, many insurance companies began excluding coverage for injury and property damage arising out of AI.\n\nFor a hardtech company that language reaches everything you do. If AI touched the work anywhere, including the tools your engineers designed it with or the product itself, the insurer can deny the claim.\n\nThe problem with these exclusions is that they get added account by account, so the same hardware can be covered at one insurer and excluded at the next, and nothing in the process makes that obvious. It shows up as a form number on the endorsement list attached to your renewal, which is why most founders don't learn they have one until they file a claim.\n\n**Why now**\n\nPhysical AI is scaling faster than the insurance market underneath it. Insurance moves slower, and its first instinct with an unfamiliar risk is to exclude it and wait for loss data. Cyber went the same way: exclusions first, then specialists who understood it, then a real market. AI is at the exclusion stage now, and these companies pushing the frontier can't wait for the cycle to finish on its own. \n\n**What we do**\n\nWe handle the whole coverage process end to end.\n\nYou send us an application. We take what you're building and put it in terms an insurance company can understand and price.\n\nThen we go to market. We have access to hundreds of insurance companies, the ones that actually write this kind of risk, and we take your file to them. Most will pass. The few that want it compete on price and terms, and we keep shopping until every requirement you have is met, including the limits your customer contracts demand.\n\nWhen offers come back we read the policies line by line. If an AI exclusion is attached, we push to get it removed, or we find a company that writes the AI coverage in on purpose. Then we tell you in plain terms what you've got.\n\nWe cover the lines your company actually needs: general liability, umbrella, property, equipment in transit and on customer sites, workers' comp, professional liability, management liability, and cyber.\n\nYou never pay us. The insurance company pays us a commission out of the premium when a policy is placed, so applying costs you nothing and there's no additional fee from us at any point.\n\n**Where this goes**\n\nPhysical AI is expanding faster than any industry, and every company in it buys insurance every year for as long as it exists. The programs grow as the fleets grow.\n\nWe want to be the brokerage that grows with them: the people who know this risk best, hold the relationships with the markets that write it, and build the loss data of the future.\n\n**Why us**\n\nWe're Sam and Alex, on leave from Harvard. We were already in insurance building models when we noticed that batch mates building some of the most innovative technology in the world couldn't get covered at all. Since then we’ve landed partnerships to offer these frontier companies easier access to better policies.\n\n![uploaded image](/media/?type=post&id=110197&key=user_uploads/2666435/74c5c462-a2ff-44b8-ac6c-f8ebc7807dbd)\n\n\n**The ask**\n\nIf you're building anything on the frontier, apply at[ risklytics.ai](http://risklytics.ai). It takes a few minutes, it costs nothing, and we'll come back with what the market says.\n\nIt doesn't matter what you have in place today. If you're building, we want to get you covered.\n\nAnd if you're an insurer, MGA or wholesaler with appetite here, we want to talk:[ risklytics.ai/markets](https://www.risklytics.ai/markets/)", "url": "https://www.ycombinator.com/launches/SfN-risklytics-insurance-for-the-frontier", "votes": 53, "created_at": "2026-08-13T16:30:00.597Z" } ] }, { "id": 32682, "name": "LATO", "slug": "lato", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/78429e3a8867e93ea2df52fc422fdbd447685774.png", "website": "https://www.latolabs.io/", "all_locations": "San Francisco, CA, USA", "long_description": "The best investments come from conviction, and conviction comes from understanding a market deeply: its customers, its competitors, and where it's heading.\r\n\r\nToday that understanding is slow and expensive to build. Reaching the right people takes weeks, a commercial study costs $300k+, and each one starts from scratch.\r\n\r\nLATO does this research end-to-end, combining public data and the fund's own knowledge with first-hand interviews of a market's consumers, experts, and operators. \r\n\r\nThis research grounds a simulation of the market to predict how market participants behave.", "one_liner": "Predict how market participants behave", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1782190140, "tags": [ "B2B", "Investing", "Market Research", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/lato", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/lato.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/lato", "x_url": "https://x.com/lato_ai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/78429e3a8867e93ea2df52fc422fdbd447685774.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/3c91e12be80a54b380afd86b5bbc965ae04ff53f.png" ], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 3156352, "full_name": "Tymek Staniszewski", "title": "Founder/CEO", "founder_bio": "Co-Founder at LATO - we predict how market participants behave\nPreviously 6 years in private equity at Verdane ($11B AUM)", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/tymeks/", "x_url": "https://x.com/tymek_st", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c870771d71e879a05e6abbc4166f0e3c57153c31.jpg" }, { "user_id": 2539014, "full_name": "Tien Chu", "title": "Founder/CTO", "founder_bio": "Co-Founder at LATO (S26), the agent-native research and simulation platform for investors. Previously ML and SWE across enterprise R&D, research labs, and early-stage startups. Studied CS, Math, and Philosophy at University of Warsaw.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/chutien/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fde5a6638eca9fd615155c8f35f1e056b8b97cb2.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110614, "title": "LATO - Research any company or market in days", "tagline": "LATO enables private equity investors to do commercial due diligence in days", "body": "[ ](https://youtu.be/1I4wuRJqanE)\nThis year alone, private equity deployed over $1 trillion. To get conviction on a deal, investors run commercial diligence - it costs c.$500k and takes weeks.\\\n\\\nOur agents source and interview hundreds of experts in any language at scale.\\\nThey combine this expert knowledge with public and proprietary data to deliver fully traceable reports.\\\n\\\nWe already work with leading funds that push the frontier.\n\nIf you are spending too much time and capital on this - schedule a demo at: \\\n\\\nWe'd love to show you how the future looks.", "url": "https://www.ycombinator.com/launches/Sm6-lato-research-any-company-or-market-in-days", "votes": 20, "created_at": "2026-08-17T19:00:00.215Z" } ] }, { "id": 32716, "name": "Neuromorphic", "slug": "neuromorphic", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/470f3b57cd018aad9c127e64e5b952d32a64d435.png", "website": "https://www.neuromorphic.vision/", "all_locations": "San Francisco, CA, USA", "long_description": "We are building robots that can work autonomously in existing wet labs. Unlike traditional lab automation, which requires dedicated equipment and fixed workflows, our robot moves between workstations, manipulates existing scientific equipment, transports lab materials, and executes experiments using the same lab infrastructure as a human technician. We make the robots easy to manage through natural interfaces like Slack, email, and phone. Our goal is to make deploying robots in wet labs as easy as onboarding a new employee, and to accelerate scientific research!", "one_liner": "Building robots to automate wetlabs", "team_size": 3, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1786056544, "tags": [ "Hard Tech", "Robotics", "B2B", "Biotech", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/neuromorphic", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/neuromorphic.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/neuromorphic-vision", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/470f3b57cd018aad9c127e64e5b952d32a64d435.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3607343, "full_name": "Vatan Aksoy Tezer", "title": "Founder/CEO", "founder_bio": "Founder at Neuromorphic. Previously Lead Robotics Engineer at Augmentus, deploying motion planning algorithms to production with sub-mm accuracy and leading a team of 5. Before that, built navigation stacks from scratch at KABAM Robotics (deployed 100+ robots) and scaled ASRS systems at Rapyuta Robotics (1000+ robots to production). Core maintainer of MoveIt at PickNik Robotics. MS in CS from Sabanci University, BS in Astronautical Engineering from Istanbul Technical University.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/vatanaksoytezer/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d22faa66df9b9d4a1f2250b01fe1287fe3e28347.jpg" }, { "user_id": 2794176, "full_name": "Ege Doganay", "title": "Founder/CTO", "founder_bio": "Founder at Neuromorphic (YC S26). Robotics and AI research at Bilkent University. Research fellow at Nanyang Technological University. Published research in top AI and robotics conferences. Won HackMIT, Stanford XR, MIT Reality Hack.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ege-doganay/", "x_url": "https://x.com/_doganay_ege", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/dda0ccf712da9c5904e6a14ea227c1d7210f36b9.jpg" }, { "user_id": 3316337, "full_name": "Cem Toker", "title": "Founder/COO", "founder_bio": "Founder at Neuromorphic. Co-founded ErgoFlow, building LLM routing and training acceleration. Previously developed static calibration algorithms for industrial robots at Zeeko achieving 30% higher precision. Engineering at Cambridge. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/cem-seref-toker", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/938e49a805b6e2d5c1ed6645ef6db6e8f54dcb9d.jpg" } ], "jobs": [ { "id": 112214, "title": "Forward Deployed Engineer", "url": "https://www.ycombinator.com/companies/neuromorphic/jobs/eae51Em-forward-deployed-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Robotics", "salary_range": "$100K - $150K", "equity_range": "0.50% - 1.50%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [ "Python", "CAD Design" ] } ], "news": [], "launches": [ { "id": 109550, "title": "Neuromorphic - Building an embodiment-agnostic “brain” that makes robots fast to deploy", "tagline": "We make robots smarter, and faster to deploy. Starting with wet labs!", "body": "Hi everyone! We’re Cem, Ege, and Vatan, the founders of Neuromorphic 👋\\\n\\\n**TL;DR** Neuromorphic is building an embodiment-agnostic “brain” that makes robots fast to deploy, and easy to manage through natural interfaces like Slack, email, and phone. Our goal is to make deploying robots in wet labs and other physical industries as easy as onboarding a new employee.\n\n\\\n\\\n**The Problem**\\\nTo make robotics mainstream, we believe two problems have to be resolved:\\\n1\\. Deployment: Robots are still too slow, expensive, and engineering-intensive to deploy and adapt to new tasks and environments.\\\n2\\. Interaction: Managing and controlling robots remains too technical.\\\n\\\n**What We’ve Built**\\\nNeuromorphic packages sensors, compute, and software into a “brain” that can be attached to any robot body. An onboard LLM orchestrator turns natural-language instructions into workflows made up of reusable robotic skills, such as navigating, picking up an object, opening a refrigerator, or inspecting an area. Individual skills can be of various types; they could be AI models we develop, like Vision-Language Models and Video Action Models, or just plain code. Sensor-based safety loops monitor every action to ensure reliable execution.\n\nWe are also giving robots emails, phone numbers, and Slack access to make human-robot interaction more intuitive than using dashboards. The robots then respond to queries and take actions using the onboard LLM.\\\n\\\n**Traction**\\\nWithin a week of our first conversation with a leading biotechnology company, we deployed our robot in their wet lab, which is capable of operating autonomously 24/7, and already completed over 1500 tasks.\\\n\\\n**The Team**\n\n![uploaded image](/media/?type=post&id=109550&key=user_uploads/3607343/b28b023b-f328-4c5d-9181-720e96e5b1a3)\n\n\\\nCem: Second-time founder and did robotics research at Cambridge University. Previously built LLM routing and training acceleration, and did precision robotics.\\\n\\\nVatan: Maintained multiple open-source robotics libraries, most notably MoveIt, one of the most popular motion-planning frameworks for robotics. Led various teams in robotics startups from Series A to Series C in Japan and Singapore. Deployed over a thousand robots.\\\n\\\nEge: Did Robotics and AI research at Bilkent University. Research fellow at Nanyang Technological University. Published research in top AI and robotics conferences. Won HackMIT in 2024.\\\n\\\nCem and Ege first met as teammates in the FIRST Robotics Competition. They reached out to Vatan for mentorship and robotics advice. We spent the next three years building robots and winning competitions together. We remained close and eventually reunited to start Neuromorphic!\\\n\\\n**Ask**\\\nWe're looking for warm intros to wet lab, biotech, and physical-industry operators. If you have a repetitive physical process and have wondered whether a robot could do it, email us at [founders@neuromorphic.vision](mailto:founders@neuromorphic.vision) or [book a call here](https://cal.com/team/neuromorphic/meet-the-founders-of-neuromorphic).", "url": "https://www.ycombinator.com/launches/SUw-neuromorphic-building-an-embodiment-agnostic-brain-that-makes-robots-fast-to-deploy", "votes": 73, "created_at": "2026-08-10T15:00:00.364Z" } ] }, { "id": 32717, "name": "Neuron Industries", "slug": "neuron-industries", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/12fc3438894b5a05c8d3d1997a329d2c5053c9b8.png", "website": "https://neuronindustries.com", "all_locations": "Los Angeles, CA, USA", "long_description": "Neuron is crafting the world’s first purpose-built industrial controller for AI, replacing the antiquated PLCs running every machine and factory. We help customers from chemical processing to nuclear reactors compress controls projects from weeks to hours. ", "one_liner": "Industrial Controllers built for AI", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1784835700, "tags": [ "Robotic Process Automation", "Manufacturing", "Automation", "Industrial" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/neuron-industries", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/neuron-industries.json", "year_founded": 2026, "city": "Los Angeles", "country": "US", "linkedin_url": "https://www.linkedin.com/company/neuron-industries-1", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/12fc3438894b5a05c8d3d1997a329d2c5053c9b8.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3676531, "full_name": "Dennis Ren", "title": "Founder/CEO", "founder_bio": "CEO and Co-Founder at Neuron Industries. Previously founding EE at Freeform, built its hardware tech stack from the ground up and scaled the world's most advanced metal 3D printing factory to production. Vision Pro System EE at Apple, owned eye tracking and Optic ID subsystems form factor system and board design. Ex-Tesla (Fremont) , Amazon (Go), Apple (iPad Pro, Pencil). UCSD BSEE.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dennisren/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/da38da0d216a741391a7f9784585f65188573dcf.jpg" }, { "user_id": 3732370, "full_name": "Kenneth Rhee", "title": "Founder/CTO", "founder_bio": "CTO and Co-Founder at Neuron Industries. Previously founded WorkIPO, bootstrapping it as a solo founder, building mission-critical enterprise software systems. Seller-marketplace systems at GOAT.com transacting hundreds millions of dollars, engineering at Amazon on Elastic Block Store (trillions I/O ops daily), Silk Browser for FireTV at Amazon, Dassault Systemes' Medidata on regulated clinical services. Computer Science at UC Irvine.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/kennethrhee/", "x_url": "https://x.com/kenneth_rhee", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/44092f707a68aa3f8c2cff8e697c1a3f2b86ebef.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107212, "title": "Neuron Industries: The world's first purpose built industrial controller for AI", "tagline": "Throw away your PLCs!", "body": "\n\n**TL;DR:** Neuron is building the world's first purpose built industrial controller for AI, replacing the antiquated PLCs running every machine and factory. We compress controls projects from weeks to hours, from chemical processing to nuclear reactors.\n\nWe have hardware ready to deploy today, hit us up for a pilot!\n\n**Problem**\n\nEvery piece of industrial equipment, conveyors, pumps, packaging cells, process plants, runs on PLCs. But PLCs got stuck in the 90s: it's 2026 and you're paying $5000 for a controller with 5MB of memory, programmable only with esoteric IDEs and languages frozen in 1993. Reindustrialization and data center buildups are pushing demand up while the programmer workforce shrinks. Projects still cost a fortune and take months.\n\n**Solution**\n\nWe're rebuilding the controller from first principles: state-of-the-art hardware, an AI-native stack, one ecosystem unifying PLC, HMI, SCADA, and Historian. Our agent contextualizes your requirements and hardware, then lets anyone program production applications in plain English, with test and verification guardrails and first-class data observability.\n\n**Traction**\n\nOur tech is in customers' plants today: pilots in chemical processing, electronics manufacturing, wastewater treatment, and small modular reactors.\n\n**Team**\n\nDennis Ren (CEO): founding EE at Freeform, built the hardware stack for the world's most advanced metal 3D-printing factory. Apple Vision Pro founding team, Amazon Go, Tesla Fremont.\n\nKenneth Rhee (CTO): second-time founder, bootstrapped and exited WorkIPO. Custom ERP, MES, and CRM systems; ex-Amazon and GOAT Group (YC W11).\n\n**Our ask**\n\nIf you — or someone you know — build machines or factories that run on PLCs and HMIs, we'd love to talk: [hello@neuronindustries.com](mailto:hello@neuronindustries.com)", "url": "https://www.ycombinator.com/launches/RtE-neuron-industries-the-world-s-first-purpose-built-industrial-controller-for-ai", "votes": 12, "created_at": "2026-07-24T18:02:58.034Z" } ] }, { "id": 32748, "name": "PokerClubHub", "slug": "pokerclubhub", "former_names": [ "Rounder" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/bdce623e4f66c5199feb962395591af2b0973e32.png", "website": "https://www.pokerclubhub.com", "all_locations": "Jaco, Puntarenas Province, Costa Rica", "long_description": "TL;DR: PokerClubHub is the first licensed poker club platform. We protect players’ deposits and allow anyone to open their own poker club and make money.\r\n\r\nRegulated online poker sites like PokerStars have open player pools and thus are full of pros and bots, driving players to unregulated private clubs.\r\n\r\nExisting poker club platforms like ClubGG have no payments processing. Players deposit with a club owner by sending them crypto for play-money chips. In theory, these chips can be exchanged back for money. However, players can get scammed and withdrawing takes ages.\r\n\r\nThis model is also broken for honest club owners who have to manually process hundreds of deposits a day. They lose out on rake when they sleep and struggle to get players to trust them.\r\n\r\nOn PokerClubhub, players deposit with PokerClubHub, not the club owner. We handle all deposits and withdrawals for the club owner, saving them time and increasing their earnings. We pay the rake (if any) to club owners as affiliate fees.\r\n\r\nWe hold the same gaming license as BetHog (built by founders of FanDuel) and Nexa Poker (part of World Poker Tour). Player deposits are secured using Fireblocks.\r\n\r\nYou can also play rake-free games with your friends and acquaintances on our platform. We ensure that everyone pays up.", "one_liner": "Host a Poker Club and Make Money", "team_size": 0, "industry": "Consumer", "subindustry": "Consumer -> Gaming", "launched_at": 1785028791, "tags": [ "SaaS", "Crypto / Web3", "Gaming", "eSports" ], "tags_highlighted": [ "Crypto / Web3" ], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Consumer", "Gaming" ], "regions": [ "Costa Rica", "Latin America" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/pokerclubhub", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/pokerclubhub.json", "year_founded": 2026, "city": "Jaco", "country": "CR", "linkedin_url": "", "x_url": "https://x.com/pokerclubhub", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/bdce623e4f66c5199feb962395591af2b0973e32.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [], "jobs": [], "news": [], "launches": [ { "id": 107788, "title": "PokerClubHub - The First Licensed Online Poker Club Platform", "tagline": "Host Your Own Poker Club, Make Money", "body": "**TL;DR:** PokerClubHub is the first licensed poker club platform. We protect players’ deposits and allow anyone to open their own poker club and make money.\\\n\\\n[https://drive.google.com/file/d/15qCencYd92aBcZdMjZxykUytkduYrGIb](https://drive.google.com/file/d/15qCencYd92aBcZdMjZxykUytkduYrGIb/view?usp=sharing)\n\n**♣️ Problem**:\n\nRegulated online poker sites like PokerStars have open player pools and thus are full of pros and bots, driving players to unregulated private clubs.\n\nExisting poker club platforms like ClubGG have no payments processing. Players deposit with a club owner by sending them crypto for play-money chips. In theory, these chips can be exchanged back for money. However, players often get scammed and withdrawals are regularly delayed.\n\nThis model is also broken for honest club owners who have to manually process hundreds of deposits a day. They lose out on rake when they sleep and struggle to get players to trust them.\n\n**Solution:**\n\n**Players deposit with PokerClubHub, not the club owner.** We handle all deposits and withdrawals for the club owner, saving them time and increasing their earnings. We return the rake to club owners as affiliate fees.\n\nWe hold the same gaming license as BetHog (built by founders of FanDuel) and Nexa Poker (part of World Poker Tour). Player deposits are secured using Fireblocks.\n\nYou can also play rake-free games with your friends and acquaintances on our platform. We ensure that everyone pays up.\n\n**Our ask:**\n\nPlease play some cards on [www.pokerclubhub.com](http://www.pokerclubhub.com). Sign up using code: “LAUNCH” and receive up to $40.\n\nContact us at [support@pokerclubhub.com](mailto:support@pokerclubhub.com) if you’re a club owner and interested in starting a club on our platform.\n\nIf you know anyone in iGaming, we’d love to chat to them! ([ivan@pokerclubhub.com](mailto:ivan@pokerclubhub.com))\n\n**Why we built PokerClubHub:**\n\n1\\. To stop players from getting scammed\n\n2\\. All the value in poker is created by players but too much of it accrues to a handful of companies. We think clubs, where hosts build communities and share in the economics, are healthier for the ecosystem.\n\n![uploaded image](/media/?type=post&id=107788&key=user_uploads/2574088/85ca689f-b4f9-4d4d-b6af-afa61336a17c)\n\n", "url": "https://www.ycombinator.com/launches/S2W-pokerclubhub-the-first-licensed-online-poker-club-platform", "votes": 14, "created_at": "2026-07-27T15:27:37.799Z" } ] }, { "id": 32749, "name": "Arbital", "slug": "arbital", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/41b5fd38dcf7826f6043c80c1a7e50338b8b5cb1.png", "website": "https://www.arbital.xyz", "all_locations": "San Francisco, CA, USA", "long_description": "Arbital lets you trade perps, memes, stocks, and run market-making strategies from one place. In our first 7 months, 2,700 traders have traded $1.35B in volume through Arbital, and we're now doing $435M/month. \r\n\r\nWe're building towards a social, mobile-first trading app that brings tokenized equities, crypto, and perps to everyone.", "one_liner": "The social trading app for stocks, memes, and perps", "team_size": 3, "industry": "Fintech", "subindustry": "Fintech -> Consumer Finance", "launched_at": 1785892617, "tags": [ "Fintech", "Crypto / Web3", "Investing", "Trading", "AI" ], "tags_highlighted": [ "Crypto / Web3" ], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech", "Consumer Finance" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/arbital", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/arbital.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/arbitalxyz/", "x_url": "https://x.com/arbital_xyz", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/41b5fd38dcf7826f6043c80c1a7e50338b8b5cb1.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 531306, "full_name": "Karon Pangestu", "title": "Founder/CEO", "founder_bio": "Co-founder at Arbital. Head of Growth and first non-engineering hire at Aevo, where I led growth, marketing, and product - helping scale the exchange from $1B to ~$100B in cumulative trading volume, and growing DAU from 1K to 30K in five months. Helped take the company public via token launch (token now listed on Binance and Bybit, ~$3.8B ATH valuation). Previously co-founded Launchy, a crypto media platform reaching 20K+ builders and investors. BA in Mathematics & Statistics, LSE.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/pangestukaron/", "x_url": "https://x.com/pangestu_karon", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3dd6c9bdc356a470d65fdf0060c83396b8b179ca.jpg" }, { "user_id": 3674258, "full_name": "Steven Mulyawan", "title": "Founder/COO", "founder_bio": "Co-founder at Arbital. Led BD and partnerships at Aevo across Asia, personally building and owning relationships with major KOLs, agencies, and high-net-worth traders - the network we now tap directly to drive Arbital's growth. Onboarded 10,000+ traders during Aevo's scale from $1B to ~$100B in cumulative volume.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/steven-mulyawan-974ab7252?utm_source=share_via&utm_content=profile&utm_medium=member_ios", "x_url": "https://x.com/0xstevezilla", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/07422eb4413225536a4e713e23fb4cd54da5d006.jpg" }, { "user_id": 3492152, "full_name": "Vincent Owen Toniwan", "title": "Founder/CTO", "founder_bio": "Co-founder at Arbital. Built and ran the backend and market-making infrastructure that powered Aevo as it scaled from $1B to ~$100B in cumulative volume - driving latency down to <2ms at 100k events/s. Built Aevo Degen from scratch, one of the first exchanges offering up to 1,000x leverage. Previously led backend at N1 (terminal.trade) and MEV infrastructure at Bitwyre.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/vincent-owen-202a9821a/", "x_url": "https://x.com/0xvincent_eth", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e93aee4429cb608bc5224c9c3761ee549666ed47.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110390, "title": "Arbital - Trade memes, stocks, and perps from one place", "tagline": "Trade Anything, Anywhere - 2,700 traders have traded $1.35B in 7 months", "body": "**TL;DR:** Arbital is a trading terminal for memecoins, stocks, and perps. In our first seven months, 2,700 traders have traded $1.35B through Arbital. We recently hit a record $41M in daily volume and are now doing \\~$435M/month in trading volume. \\\n\\\n\n\n**The problem**\n\nVolatility no longer lives in one market.\n\nOn any given day, it might be a memecoin ripping, an AI stock pumping, or a Korean or Chinese stock moving.\n\nBut two things get in the way:\n\n* **Discovery is noisy.** Trade ideas are scattered across social feeds, and it's hard to know who's credible or why they're making a trade.\n* **Execution is fragmented.** There's no single place to trade every market that matters - crypto, memes, perps, and equities.\n\nArbital brings traders, trade ideas, and markets into one place.\n\n**What we built**\n\nArbital is one place to discover and trade the markets you care about most.\n\nWith Arbital, you can:\n\n* Trade memecoins, equities, and perps\n* Run automated strategies like market making and grid trading\n\nWe’re also building a social, mobile-first trading app where users can discover what other traders are buying, understand their reasoning, and execute immediately.\n\n**Traction - first 7 months**\n\n* **$1.35B+** total trading volume\n* **$41M** in a single day - a new all-time high\n* **\\~$435M/month** in trading volume\n* **2,700** traders\n\n**Why us**\n\nBefore Arbital, we were early team members at Aevo. We helped scale the exchange from \\~$1B to $100B in cumulative volume, grew DAU from 1,000 to 30,000 in five months, and helped launch its token, which reached a \\~$3.8B peak valuation and is listed on Binance and Bybit.\n\nOn the engineering side, we built trading systems that handled 100K events per second with <2ms latency.\n\nNow we're using what we learned to build a simpler trading experience for a global audience.\n\n![uploaded image](/media/?type=post&id=110390&key=user_uploads/531306/fe17e817-a600-47c9-8476-41dabab20359)\n\n**Our asks**\n\nIf you trade memes, perps, or stocks, we'd love for you to try Arbital and tell us what you think. Trader feedback helps us make the product better.\\\n\\\n**Try Arbital:** \n\n**Feedback / intros:** [karon@intric.org](mailto:karon@intric.org)\n\n**Telegram:** @kp_9961", "url": "https://www.ycombinator.com/launches/SiU-arbital-trade-memes-stocks-and-perps-from-one-place", "votes": 69, "created_at": "2026-08-15T02:00:00.546Z" } ] }, { "id": 32750, "name": "Ultrasonium", "slug": "ultrasonium", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/734dc12d12265ba8890bf4ef5e7e564bdb629a87.png", "website": "https://ultrasonium.com/", "all_locations": "Boston, MA, USA", "long_description": "The technologies that define this century, including data centers, rockets, satellites, and robots, are built from complex metal parts. As these systems advance, the parts get harder to make. Past a certain complexity, casting and machining stop working, and the advanced methods that remain are slow, expensive, and hazardous.\r\n\r\nUltrasonium has invented a new manufacturing method to make these critical parts 15x faster and 75% cheaper than the methods they replace, in the U.S. and at scale. Solid metal feedstock goes in and a finished part safely comes out, controlled by a physical AI layer and novel control processes. \r\n\r\nThe systems that matter most, in aerospace, energy, and compute, are demanding metal geometries that can't be machined or cast and are too expensive and slow to make with incumbent methods, at a moment when domestic manufacturing capacity is a national priority.\r\n\r\nOur founding team has built and innovated some of the most complex engineering systems known to man: quantum computers, superconductors, nuclear reactors, and novel aerospace alloy processing methods. ", "one_liner": "Metal manufacturing, faster, cheaper, and more versatile.", "team_size": 5, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1785870310, "tags": [ "Hard Tech", "Hardware", "Robotics", "Manufacturing" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/ultrasonium", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/ultrasonium.json", "year_founded": 2026, "city": "Boston", "country": "US", "linkedin_url": "https://www.linkedin.com/company/ultrasonium/", "x_url": "https://x.com/Ultrasonium", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/734dc12d12265ba8890bf4ef5e7e564bdb629a87.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3735756, "full_name": "Jack Qiu", "title": "Founder/CEO", "founder_bio": "Founder & CEO of Ultrasonium. Inventor, physicist, and engineer with a PhD in EECS from MIT, with research spanning quantum computing, cryogenics, plasma research, signal processing, hardware control, and advanced manufacturing. First author of a Nature Physics publication in collaboration with MIT Lincoln Laboratory under IARPA and DoD programs. Former researcher at Lawrence Berkeley National Laboratory. BA in Physics from UC Berkeley.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jack-q-a1b81546/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/10afaa4d3df89c18be993661d93702c07d1f3782.jpg" }, { "user_id": 3748345, "full_name": "Hunter Brown", "title": "Founder", "founder_bio": "Founder & VP, Engineering at Ultrasonium\n\nStarted my R&D career working at national labs by age 15. I have contributed to the ITER fusion reactor and CERN through labs at NHFML and LBNL. I have led teams of fabricators and designers in both research and industry. I am passionate about hard tech and moving groundbreaking research ideas into real products. \n\nI lead the Ultrasonium team on the novel and cutting edge technology we implement in our manufacturing process. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/hunterpbrown/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/525b78441a81f2ea58c6801c81cc2f8643e4e421.jpg" }, { "user_id": 1892793, "full_name": "Chris Carter", "title": "Founder", "founder_bio": "Founder & Chief Scientific Officer at Ultrasonium. PhD candidate on leave from MIT researching novel metal matrix materials and 3D printed metal heat treatments under NSF GRFP, MIT UCEM, GEM, and .406 Venture fellowship programs. Previously at Argonne National Laboratory working on 3D printed steel. Led metal melt pool simulations for NASA-funded multi-university project. Patent holder and publisher on metal printing. MS in Mechanical Engineering from MIT, BS from University of South Carolina.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/christopher-p-carter/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d2c39cf3788b615d4686293df632fc23d11aab12.jpg" }, { "user_id": 2875171, "full_name": "Alexander Urbanski", "title": "Founder", "founder_bio": "Founder & Chief Research Officer of Ultrasonium. Former founder of the DOE-funded nuclear energy startup Bay State Energy and inventor of the Nuclear Thermoelectric Generator. Studied physics and computer science at Boston University. First hire at Phase3D - built and commercialized metals manufacturing technology deployed across NASA, USAF, and companies across four continents. Former metals manufacturing research engineer at Argonne National Laboratory. Dropped out of mech-e grad program. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aurban42/", "x_url": "https://x.com/urban_alex42", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/453e1928db84e3ca68b98855a033c46f98ebc438.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 113464, "title": "Ultrasonium: Shaping the Future of Metal Manufacturing", "tagline": "We make the world’s most difficult metal parts faster, cheaper, and more versatile.", "body": "**The problem**\n\nThe technologies that define this century, including data centers, rockets, satellites, robots, etc., are built from complex metal parts. As these systems advance, the parts get harder to make. Past a certain complexity, casting and machining stop working, and the advanced methods that remain are slow, expensive, and limited in the alloys and geometries they can process.\n\n**What we built**\n\nUltrasonium has invented a new method to make these critical parts. Solid metal feedstock goes in and a finished part comes out, controlled by a Physical AI layer and novel control processes.\n\nWe produce near-net-shape complex metal parts up to 15× faster, 75% cheaper, and with far greater alloy versatility than conventional manufacturing methods, at scale in the U.S.\n\n\n\n![uploaded image](/media/?type=post&id=113464&key=user_uploads/3735756/dbd12c5d-1203-4e50-9939-980a218df45a)\n\n**Why now / why us**\n\nThe systems that matter most, in aerospace, defense, energy, and compute, are demanding metal geometries that can’t be machined or cast and are too expensive and slow to make with incumbent methods, at a moment when domestic manufacturing capacity is a national priority.\n\nOur founding team has worked on some of the most difficult engineering systems in the world, including quantum computers, superconductors, nuclear reactors, and advanced aerospace alloy processing. Between the four of us, we have backgrounds in physics, metallurgy, advanced manufacturing, controls, hardware, and commercialization.\n\nJack Qiu (CEO) completed his PhD in Electrical Engineering and Computer Science at MIT and published pioneering quantum computing work in Nature Physics. Chris Carter (CSO) left to join Ultrasonium during the 5th year of his PhD program at MIT after more than six years of research in state-of-the-art metallurgy and materials processing. Alexander Urbanski (Chief Research Officer) previously founded a DOE-funded nuclear startup and was the first hire at Phase3D, where he built and commercialized metal manufacturing monitoring technology. Hunter Brown (VP, Engineering) began working at national labs when he was 15, studied physics at UC Berkeley, and contributed to the ITER fusion reactor.", "url": "https://www.ycombinator.com/launches/TW4-ultrasonium-shaping-the-future-of-metal-manufacturing", "votes": 6, "created_at": "2026-09-02T19:54:37.419Z" }, { "id": 109239, "title": "Ultrasonium: A new way to make complex metal parts, 15x faster and 75% cheaper", "tagline": "Ultrasonium makes the most difficult metal parts for the world's most important systems 15x faster and 75% cheaper. ", "body": "\n**The problem**\n\nThe technologies that define this century - data centers, rockets, satellites, robots, etc. - are built from complex metal parts. As these systems advance, the parts get harder to make. Past a certain complexity, casting and machining stop working, and the advanced methods that remain are slow, expensive, and hazardous.\n\n**What we built**\n\nUltrasonium has invented a new method to make these critical parts. Solid metal feedstock goes in and a finished part safely comes out, controlled by a physical AI layer and novel control processes. We produce near-net-shape complex metal parts 15x faster and 75% cheaper than the methods they replace, in the U.S. and at scale.\n\n \n\n![uploaded image](/media/?type=post&id=109239&key=user_uploads/2875171/55d1e111-fc43-450d-a83c-2c191d88d909)\n\n\n**Why now / why us**\n\nThe systems that matter most, in aerospace, defense, energy, and compute are demanding metal geometries that can’t be machined or cast and are too expensive and slow to make with the incumbent methods, at a moment when domestic manufacturing capacity is a national priority.\n\nOur founding team has built and innovated some of the most complex engineering systems known to man - quantum computers, superconductors, nuclear reactors, and novel aerospace alloy processing methods. Jack Qiu (CEO) completed his PhD in Electrical Engineering and Computer Science at MIT and has published pioneering quantum computing work in Nature Physics. Chris Carter (CSO) dropped out during the 5th year of his PhD program at MIT and has over a half dozen years of experience researching state of the art metallurgy and materials processing systems. Alexander Urbanski (Chief Research Officer) previously founded a DOE-funded nuclear startup and was the first hire at Phase3D where he built and commercialized metals manufacturing monitoring technology. Hunter Brown (VP, Engineering) began working at national labs when he was 15, graduated from UC Berkeley Physics, and contributed to the ITER fusion reactor.", "url": "https://www.ycombinator.com/launches/SPv-ultrasonium-a-new-way-to-make-complex-metal-parts-15x-faster-and-75-cheaper", "votes": 8, "created_at": "2026-08-05T17:31:24.990Z" } ] }, { "id": 32751, "name": "Vestris", "slug": "vestris", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/e17d90c0a32a27bdb75bac6bb8a2ca4e0ad1646d.png", "website": "https://vestris.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Vestris automates title processing end-to-end, from contract to close, so title shops can close more deals with fewer processors. Title insurance has famously high margins, but slow, manual processing is a constant pain — especially in today's bleeding housing market.", "one_liner": "AI native platform for real estate closings", "team_size": 2, "industry": "Real Estate and Construction", "subindustry": "Real Estate and Construction -> Housing and Real Estate", "launched_at": 1783395705, "tags": [ "Workflow Automation", "Home Services" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Real Estate and Construction", "Housing and Real Estate" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/vestris", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/vestris.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/119274035", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/e17d90c0a32a27bdb75bac6bb8a2ca4e0ad1646d.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/84c006c8048ba11d7ef2471745824fa60204e6d4.png" ], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3411757, "full_name": "Joshua Tang", "title": "Founder", "founder_bio": "Founder at Vestris, with the goal of becoming an AI-native title company. Studied Computer Engineering at the University of Michigan, and built telemetry software for MRacing, the school's FSAE team.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/joshua-tang-449717279/", "x_url": "https://x.com/joshtang07", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/075afc25dd0ae0479f01d13c1b8edf88767414fc.jpg" }, { "user_id": 3411947, "full_name": "Aahil Valliani", "title": "Founder", "founder_bio": "Founder at Vestris. Previously founded Safe Kids AI, creating better mental health software for K-12 schools and scaled to over 10k+ users. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aahil-valliani", "x_url": "https://x.com/aahil__val", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b1b1692674aac1dc411abf6cae40e4deccd6f5f1.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109429, "title": "Vestris: Al native title processing service", "tagline": "We allow title shops to close more deals", "body": "**TL;DR:** Title insurance is the thorn in the side of every single residential real estate transaction in America. With the home market bleeding, you'll find attorneys processing documents by hand with outdated software at title agencies. Vestris automates the back office and supplements processors the way software should — _letting shops give the same care to closing the biggest purchase of a buyer's life._\n\n\n\n### The Problem\n\nImagine you're shopping the market and you just found your dream home. You're ready to purchase, the contract is signed, everything is in place.\n\nAnd you're recommended a settlement company, and told to buy title insurance.\n\nMost homeowners couldn't tell you what title insurance actually buys them. In a nutshell: **you're paying for the peace of mind that the property you're buying has no open claims against it** — payments a title company makes on your behalf.\n\nBut it's also a process that takes weeks. Behind the phone, you've got lawyers processing files, manually checking emails, and scanning two seemingly identical documents for changes.\n\nMeanwhile, the incumbent software is a decade-old, per-file database whose premium plans put it out of reach for the small and mid-size shops that make up most of the market.\n\n### The Solution\n\nVestris runs the file end to end:\n\n**Reads everything on arrival** — every document that hits a processor's inbox is routed to the right order, new contracts spin up new files, and the routine back-and-forth is handled, so the first 30% of a file is done before a human even looks.\n\n**Does the math** — logs into county tax and land portals to pull up-to-date figures, then prorates taxes, utilities, HOA dues, and special assessments to the settlement date on that county's convention, every figure traced back to its source.\n\n**Drafts the commitment** — building Schedule A from the vesting and legal description and Schedule B from the exceptions in the record, the work a small shop waits a week and pays an abstractor hundreds for.\n\n**Closes it out** — reconciling the closing disclosure against the settlement statement, handling the lender back-and-forth, and showing the processor exactly what changed, so no one's squinting at two seemingly identical PDFs.\n\nEvery step still goes through human approval. What changes is bandwidth: **a mom-and-pop shop never has to turn down an order because they're swamped**, and they get to spend their time on the in-person care that made buyers seek them out in the first place.\n\n### The Team\n\nWe're Aahil and Josh, best friends and UMich roommates.\n\nAahil previously built SafeKids and holds three USPTO patents. He spent the summer closing files at a title shop, seeing the processor's grind firsthand.\n\nJosh previously built live-telemetry and communication systems for Michigan's Formula SAE team. He worked the same title desk, and stole the cubicle with the monitor 😠\n\n![uploaded image](/media/?type=post&id=109429&key=user_uploads/3411947/af48be0b-23b2-4cb5-9c1c-4eb5c7b45e50)\n\n### Our Ask\n\nWe're running Vestris in parallel at a title shop that settles millions in deals a month.\n\nIf you know anyone at **a title agency, escrow, or settlement company** — we'd love an intro.\n\nAnd if you've ever bought or sold a home, reach out! You probably have strong feelings about the title process :)\n\nYou can reach us at [founders@vestris.ai](mailto:founders@vestris.ai), or better yet, give us a call at +1 (571) 466 7878.", "url": "https://www.ycombinator.com/launches/SSz-vestris-al-native-title-processing-service", "votes": 15, "created_at": "2026-08-06T19:00:00.169Z" } ] }, { "id": 32752, "name": "Moving Atoms", "slug": "moving-atoms", "former_names": [ "Run Lobster", "RunLobster", "Demon AI", "RunDemon", "Avicenna Care", "Agent Insurance" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ee4bd602221f77fc05dda940b7202523dd2b9d10.png", "website": "https://movingatoms.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Our first model, Atom 1, pre-trained on the internets video, and fine tuned for physics. \r\nIt has been action conditioned, so you can simulate robot actions within it.\r\nThis lets Robot Labs train their robots using GPUs instead of relying on human collected manual data that is slow and expensive.\r\nIn unseen and hard to replicate environments, this is the only path to getting ANY training data.", "one_liner": "Virtual Reality for Training Robots", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1786481200, "tags": [ "Hard Tech", "Robotics", "Virtual Reality", "Data Labeling", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/moving-atoms", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/moving-atoms.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ee4bd602221f77fc05dda940b7202523dd2b9d10.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 1776350, "full_name": "Shehryar Saroya", "title": "Founder/CEO", "founder_bio": "Previously AI @ TikTok (Core Recomendation Algorithm). \nMaths @ Oxford.", "is_active": true, "linkedin_url": "https://linkedin.com/in/shehryar-saroya", "x_url": "https://x.com/SaroyaShehryar", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/eba8f18828b5f957663aea131e4a12ec89f9afbb.jpg" }, { "user_id": 2375000, "full_name": "Ahmad Saroya", "title": "Founder", "founder_bio": "Studied Mathematics @ Oxford, Trained Neural nets for GCHQ\nBritish Mathematical Olympiad Gold Medalist. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ahmad-saroya-a05400237/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ab9fc0f9df4310e5df12d2db3ff69d64e603d611.jpg" } ], "jobs": [], "news": [], "launches": [] }, { "id": 32753, "name": "Ekho Labs", "slug": "ekho-labs", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/8de6c5bf64155f5dfde0293d76560b38b3e3b46f.png", "website": "https://ekholabs.com/", "all_locations": "New York City, NY, USA", "long_description": "Ekho predicts the wars, strikes and tariffs that break global shipping, and tells freight teams which shipments are exposed and what to do about them. Predicting escalation is what our team did for eight years before this - now we bring the same engine to logistics.", "one_liner": "World Model for Freight", "team_size": 0, "industry": "B2B", "subindustry": "B2B -> Supply Chain and Logistics", "launched_at": 1785301746, "tags": [ "Hard Tech", "SaaS", "Supply Chain", "Enterprise Software", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Supply Chain and Logistics" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/ekho-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/ekho-labs.json", "year_founded": null, "city": "New York City", "country": "US", "linkedin_url": "https://www.linkedin.com/company/ekho-labs", "x_url": "https://x.com/ekholabs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/8de6c5bf64155f5dfde0293d76560b38b3e3b46f.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 3683165, "full_name": "Alina Park", "title": "Founder/CEO", "founder_bio": "Founder at Ekho Labs. Previously Founding Engineer at Inca Digital, where she built a DARPA-funded AML platform, deployed across DoD, CFTC, PayPal, and more. \n\nAwarded Innovator of the Year by CFTC, prev. built Yachay AI, achieving state-of-the-art text-based geolocation (30km median error, 80+ languages). \n\nGuest speaker at Columbia MBA and Stanford Engineering.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/alinapark/", "x_url": "https://x.com/olenyapark", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9565da5bd46340e43022e967896174e95bfc8264.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109994, "title": "Ekho Labs - Forecasting Supply Chain Disruptions", "tagline": "Decision Engine for Supply Chains", "body": "**TLDR:**\n\n**Companies with global supply chains lose money to every small disruption. Ekho makes sure they don't.**\n\n**Problem:**\n\nManufacturing means choosing suppliers, booking transport, and keeping pulse on new tariffs - to make sure your production is fine. Companies tend to find out about disruptions when things break and cost them millions.\n\n**Solution:**\n\nEkho customers get an early warning AND optimal solution 1-3 weeks ahead of the events. So ships keep moving, production lines keep running, and customer promises get kept.\n\nI used to build signal detection systems for DARPA. Now that same type of engine watches 1M+ sources for the industry our phones, food, and clothes depend on - Supply Chains. What you get:\n\n* **Freight forwarders:** container-level forecasts. We’ll tell you what happens, when, and what to do while freight is still available and cheap. Your clients hear about the disruption from you, not from the news.\n* **Manufacturers and retailers:** you know which inbound materials will miss the production run, launch, or retailer order - and the cheapest fix, from expediting to switching suppliers or lanes.\n\n**Proof we’re cool:**\n\n1. When the Strait of Hormuz closed in February, we correctly predicted ceasefire dates, duration, and next steps in the war ()\n2. We already work with large global companies\n\n**Ask**\n\nIntro us to Supply Chains VPs/COO at a retail (Target, Walmart, etc.), manufacturing (Fashion, Electronics, Food, etc.), or Freight Forwarding companies.\n\n**Email:** [alina@ekholabs.com](mailto:alina@ekholabs.com)\n\n**Website:** \n\nMy hilarious twitter: ", "url": "https://www.ycombinator.com/launches/Sc6-ekho-labs-forecasting-supply-chain-disruptions", "votes": 6, "created_at": "2026-08-11T21:00:00.268Z" } ] }, { "id": 32781, "name": "Isengard Industries Inc", "slug": "isengard-industries-inc", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/17f019cfd01fa25a85e88f6a43f39d133e5ecaff.png", "website": "http://isengardindustries.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Isengard Industries is not another drone company.\r\n\r\nWe build sovereign production: the means for allied nations to make precision-strike and counter-drone systems by the thousand, on their own soil, at a fraction of what the primes charge.\r\n\r\nWe call it Weapons of Mass Production. The Arsenal of Democracy, rebuilt for the drone age.\r\n\r\nMore to come, and only where it helps. You know how to reach us.", "one_liner": "Mass-produced AI strike systems and Counter-UAS", "team_size": 30, "industry": "Industrials", "subindustry": "Industrials -> Defense", "launched_at": 1782404945, "tags": [ "Swarm Robotics", "Unmanned Vehicle", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Defense" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/isengard-industries-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/isengard-industries-inc.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/isengard-industries/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/17f019cfd01fa25a85e88f6a43f39d133e5ecaff.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 3554287, "full_name": "Francisco Serra-Martins", "title": "Founder/CEO", "founder_bio": "Founder at Isengard Industries Inc. Previously co-founder/CEO at Terminal Autonomy, scaling to $60M revenue with customers including U.S. Army, NATO, and Ukrainian armed forces. Raised $15M+ from investors. Prior CEO at Sonder Design (listed on Forbes) and Combat engineer in Australian Defence Force. Studied Engineering at University of Sydney.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/franciscoserramartins/", "x_url": "https://x.com/FranciscoSerraM", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d2f29dd1ada67b914d72ae7d29c7908601e110f5.jpg" }, { "user_id": 3752974, "full_name": "James Tennant", "title": "Founder", "founder_bio": "I'm a co-founder of Isengard Industries. We build attack drones and counter-drone interceptors that allies manufacture locally at a fraction of prime cost. \n\nBefore Isengard I served as an Australian Army Officer. I have spent a decade in private equity and private credit. I ran Balmoral, a Defense Investment Bank, and have led diligence and structuring on defence-tech deals across Australia, Ukraine and the Gulf. \n\nI'm also completing a PhD on economic warfare and hybrid threats.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jamesjtennant/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/00c28e38e0f0496272a7eb753df9756a88fb9181.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 114670, "title": "Isengard Industries Inc", "tagline": "AI Weapons of Mass Production", "body": "Hi YC, I’m Francisco, co-founder of Isengard (S26). We build precision-strike and counter-drone systems that allied nations can manufacture on their own soil.\\\n\\\n\n\nWe have $10M in revenue and a $350M pipeline spanning the US Army, Ukraine, and allies globally.\n\nI previously co-founded Terminal Autonomy and scaled it to $60M in revenue in Ukraine. Just weeks ago, its Kyiv factory was hit by a Russian ballistic missiles. [The Guardian ](https://www.theguardian.com/world/2026/jul/30/us-firm-drone-factory-terminal-autonomy-ukraine-russia)article.\n\n**With my family in Ukraine, for me, resilient defense manufacturing is personal.** Isengard designs for affordable, distributed production, giving allies the capacity to build and replenish their own arsenals.\n\nWe call it Weapons of Mass Production.\n\n**Our ask:** Introductions to allied defense procurement teams and manufacturing partners. Contact [hello@isengardindustries.com](mailto:hello@isengardindustries.com).\n\n![uploaded image](/media/?type=post&id=114670&key=user_uploads/3554287/21bb2b9f-c02c-43b6-a778-84a16a1d7414)\n\n", "url": "https://www.ycombinator.com/launches/TpW-isengard-industries-inc", "votes": 13, "created_at": "2026-09-10T19:30:05.764Z" } ] }, { "id": 32782, "name": "Denta", "slug": "denta", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/8126dcec05004e9a335df14d0116cf4774a61b06.png", "website": "https://denta.com", "all_locations": "San Francisco, CA, USA", "long_description": "Starting with dental benefits, we are relentlessly simplifying the process for patients to receive better coverage at a lower cost by putting the buying power back into the hand of the individual.", "one_liner": "Fortune 500 Dental Benefits for Startups", "team_size": 1, "industry": "Fintech", "subindustry": "Fintech -> Insurance", "launched_at": 1782407143, "tags": [ "Fintech", "Healthcare", "Health Insurance", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech", "Insurance" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/denta", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/denta.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/trydenta/", "x_url": "https://x.com/jackbeecher23", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/8126dcec05004e9a335df14d0116cf4774a61b06.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1959995, "full_name": "Jack Beecher", "title": "Founder/CEO", "founder_bio": "Dartmouth dropout. Studied CS & Math. Removing the biggest problem from dentistry: insurance.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jack-beecher-459a13180", "x_url": "https://x.com/jackbeecher23", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ec1fbfdbc45b527c4f8961d1bf237f521fe2b6eb.jpg" } ], "jobs": [ { "id": 109683, "title": "Founding Growth", "url": "https://www.ycombinator.com/companies/denta/jobs/re7M7Ih-founding-growth", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "$100K - $130K", "equity_range": "1.00% - 2.00%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 111300, "title": "Denta 🦷 Dental Benefits", "tagline": "Save time and money on dental benefits", "body": "Denta makes dental benefits 25% cheaper by selling plans directly to employers and processing claims instantly with AI.\n\n**Growing up in Dentistry → Solving Its Biggest Problem**\n\nI grew up in the back office of my dad’s dental practices, spent my high school years answering patient calls, and built software for over 300 dentists during college. I'm now solving the industry's biggest problem: insurance.\n\n**The Waste**\n\nEveryone knows how bad health insurance is, dental insurance is worse. In the US, dental insurers collect $100B per year. Despite the $25B of that spent on administration, it still takes insurers three weeks to pay claims with paper checks that Denta can pay electronically in seconds.\n\n![uploaded image](/media/?type=post&id=111300&key=user_uploads/1959995/0f923bd6-9f67-40bc-bdbe-474a4882e652)\n\n**Saving you money**\n\nDenta helps startups self fund their dental insurance. Most Fortune 500s choose this model of insurance to save money while providing better employee benefits. For startups, this model was previously inaccessible.\n\n**Mission**\n\nDenta will not stop until all of America is covered. We believe by removing the friction and waste in the industry, there will be a new age of abundance where both the dentist and patient win.\n\n**Insure like the best**\n\nIf you want to save money on dental benefits while providing the best possible experience for your employees, sign up at [denta.com](http://denta.com)", "url": "https://www.ycombinator.com/launches/SxA-denta-dental-benefits", "votes": 8, "created_at": "2026-08-26T21:00:00.368Z" } ] }, { "id": 32814, "name": "Computable", "slug": "computable", "former_names": [ "Hobby.Trade" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/4778fcb121d43a885f6b39a13d66a2fe2364fb1a.png", "website": "https://getcomputable.com", "all_locations": "New York City, NY, USA", "long_description": "", "one_liner": "Derivatives Exchange for Compute", "team_size": 2, "industry": "Fintech", "subindustry": "Fintech", "launched_at": 1784612730, "tags": [ "Marketplace", "Finance", "Cloud Computing", "Trading", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/computable", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/computable.json", "year_founded": null, "city": "New York City", "country": "US", "linkedin_url": "", "x_url": "https://x.com/getcomputable", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/4778fcb121d43a885f6b39a13d66a2fe2364fb1a.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [], "jobs": [], "news": [], "launches": [ { "id": 110494, "title": "Computable: The Exchange for Physical Compute", "tagline": "the first exchange for physical compute where you can buy, sell, use, and hedge GPU capacity for exactly the dates you want", "body": "Hi, we're Computable, the first exchange for physical compute where you can buy, sell, and use GPU capacity for exactly the dates you want like staying on Airbnb. You pay less, commit less, and stay flexible.\n\n\n\n**The problem**\n\nCompute is the newest major commodity, but it still trades like an asset from the last century. You overpay going in and you eat the weeks you don't use.\n\n**How it works**\n\nWe put a public price on every week. Open the market and you'll see dedicated H100 nodes, every week through January 2027, each with its own price, from $2.55/GPU-hr. Pick a start week and an end week, and it's yours.\n\n* **Buy the exact weeks you need.** Two nodes for the last two weeks of October is just a few clicks, and you never pay for weeks you don't want.\n* **Sell back what you don't use.** If your plans change, sell the weeks back to the market whenever you want; we keep a market quote posted on every position.\n* **Lock next year at a price you can see.** January 2027 is on the grid today, so you know exactly what the future run costs before another H100 40% rate jump happens.\n* **Real machines, fast.** Dedicated H100 SXM nodes, full SSH root access, live within an hour of your window start.\n\n**Why us**\n\nOur founding team comes from Jump Trading, Coinbase, Citadel Securities, and other trading firms and exchanges. We have spent our careers building markets. Stay tuned for more products coming soon as we build out the financial infrastructure of the AI economy.\n\n**Our asks**\n\n* **Open [market.getcomputable.com](http://market.getcomputable.com) and buy GPU!**\n* **Tell us what to list next: GPU type, region, dates.** Email [team@getcomputable.com](mailto:ray@getcomputable.com); it directly sets what goes on the grid.\n* **If GPU prices move your margins or you are worried about your long-term contracts, email [team@getcomputable.com](mailto:ray@getcomputable.com).** We have a super cool product in pipeline for exactly this. Early testers get it first!", "url": "https://www.ycombinator.com/launches/SkA-computable-the-exchange-for-physical-compute", "votes": 26, "created_at": "2026-08-18T19:00:00.333Z" } ] }, { "id": 32815, "name": "TryNearby", "slug": "trynearby", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/043d0df2ce658d03b7ad1d70bb0deaa3f2282ea5.png", "website": "https://trynearby.com", "all_locations": "Los Angeles, CA, USA", "long_description": "TryNearby connects local businesses with creators who live nearby. Restaurants subscribe monthly and automatically get booked visits from local creators who film, edit, and post a review to the local algorithm. This creates a steady stream of content that compounds into discovery on TikTok, Instagram, and increasingly AI search. Over 100 restaurants and thousands of creators are on the platform across Southern California, with AI agents handling matching, booking, and every creator conversation.", "one_liner": "Local creators for local businesses", "team_size": 5, "industry": "B2B", "subindustry": "B2B -> Marketing", "launched_at": 1786488480, "tags": [ "Restaurant Tech", "Marketing", "Creator Economy", "SMB" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Marketing" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/trynearby", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/trynearby.json", "year_founded": 2025, "city": "Los Angeles", "country": "US", "linkedin_url": "https://www.linkedin.com/company/trynearby", "x_url": "https://www.x.com/trynearby_", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/043d0df2ce658d03b7ad1d70bb0deaa3f2282ea5.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3490413, "full_name": "Yousef Abdelfattah", "title": "Founder/CEO", "founder_bio": "Co-founded FaZe Clan after starting as a 14 year old YouTuber in 2010. Spent the last 15 years in the creator economy turning FaZe into the first internet-native content company to IPO at $1B. Now building TryNearby; for local creators and businesses.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/yousef-abdelfattah-b0906b336/", "x_url": "https://x.com/fazeapex", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/52fe637c6c284fe53525e44b365bc3238f135ff9.jpg" }, { "user_id": 426164, "full_name": "Obaida Albaroudi", "title": "Founder/COO", "founder_bio": "Co-Founder at TryNearby. Early Replit employee and former Goldman Sachs. Cofounder of Clipping Exe, one of the first and largest video clipping infrastructures, processing tens of thousands of videos per month.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/obaida-a-b7b366279", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a29f3fda4d88efcf96283b58215738c69a92286c.jpg" }, { "user_id": 2518184, "full_name": "Ahmad Ibrahim", "title": "Founder/CTO", "founder_bio": "Founder at TryNearby. Previously CTO/Cofounder at Clipping Exe, building video clipping infrastructure that scaled to tens of thousands of videos monthly.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ahmibrahim97/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d5b633fb361c59514c234a4c2f456173f7dc26d6.jpg" } ], "jobs": [ { "id": 111232, "title": "Harness Engineer", "url": "https://www.ycombinator.com/companies/trynearby/jobs/MtlKiOm-harness-engineer", "location": "CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$120K - $150K", "equity_range": "0.25% - 1.00%", "experience": "1+ years", "visa": "US citizen/visa only", "skills": [ "AI Agents" ] }, { "id": 111234, "title": "Founding Operations Lead", "url": "https://www.ycombinator.com/companies/trynearby/jobs/LUvHWP0-founding-operations-lead", "location": "CA, US", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "$100K - $130K", "equity_range": "0.25% - 0.75%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 111236, "title": "Founding Sales/Growth Lead", "url": "https://www.ycombinator.com/companies/trynearby/jobs/Q8SAxNF-founding-sales-growth-lead", "location": "CA, US", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "$80K - $120K", "equity_range": "0.10% - 0.50%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 111457, "title": "Founding Content Lead", "url": "https://www.ycombinator.com/companies/trynearby/jobs/696hTHA-founding-content-lead", "location": "CA, US", "type": "Full-time", "role": "Marketing", "role_type": null, "salary_range": "$120K - $150K", "equity_range": "0.25% - 1.00%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 111459, "title": "Founding Creator & Community Lead", "url": "https://www.ycombinator.com/companies/trynearby/jobs/Eoisuz0-founding-creator-community-lead", "location": "CA, US", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "$100K - $130K", "equity_range": "0.25% - 0.75%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 110956, "title": "TryNearby: Every local business needs to be discoverable. That $150B market is moving toward local creators.", "tagline": "Word of mouth at scale for local restaurants", "body": "Hey everyone,\n\nWe're Yousef, Obaida, and Ahmad, the founders of **TryNearby**.\n\nYousef co-founded FaZe Clan and spent 15 years creating content for 10M+ followers alongside hundreds of other creators. Ahmad and Obaida co-founded one of the first social media clipping companies, building the infrastructure that tracked billions of views across thousands of videos for creator campaigns. We've watched creator content become the most powerful distribution channel on the internet, and we're bringing it to local businesses.\n\n## **TL;DR**\n\n84% of small businesses on TikTok say it helped grow their business. Almost every restaurant owner we meet knows that and has no idea how to act on it.\n\nTryNearby handles it. Every month we send a new group of creators who live nearby to visit, film, and post a review, for a $399 subscription. All of it is handled by us and our AI agents.\n\nToday, 120+ restaurants are subscribed and have had 2,500+ videos posted for them. We nearly doubled one restaurant's revenue in their first month, and owners tell us people walk in every day saying \"we saw you on TikTok.\" Almost half our new restaurants now come from other owners referring us.\n\nOur launch video: \n\n## **The problem**\n\nThere are about 1 million restaurants and cafes in the US. Most of them are invisible in the places people actually decide where to eat.\n\nThe options an owner has today:\n\n* Agencies cost thousands a month and often don’t drive real results\n* Influencers work, but finding the right local ones, negotiating, and doing it again every month is a job by itself\n* Creating their own content is a second full-time job they don't have time for, or a full-time hire they don't have the budget for\n\nAnd here's the thing: _the algorithm doesn't reward a business talking about itself. It rewards people talking about the business._\n\nSo the thing that actually works — someone who lives nearby posting about your food — happens once or twice, or not at all.\n\nAnd the platforms keep rewarding the ones who show up. TikTok's new Local Feed ranks by location and recency, so the restaurants getting fresh content every month reach their own neighborhood. The ones who don't, disappear.\n\n## **The solution**\n\nTryNearby is a subscription that puts local creators inside local restaurants every month, and their followers through the door.\n\n$399/month. Every month we automatically schedule a new group of vetted creators who live nearby. They come in, film, edit, and post. The restaurant just serves the food.\n\n![uploaded image](/media/?type=post&id=110956&key=user_uploads/3490413/7721f4f4-96e4-4a68-b860-841fee782765)\n\n### **Part 1 — They sign up, creators start showing up**\n\nThe moment a restaurant joins, our matching agent finds the creators nearby who fit that spot and sends the invites. Sometimes a creator is scheduled to visit within five minutes of a restaurant signing up. Creators pick their own time from the restaurant's availability. The restaurant doesn't coordinate anything or answer a single message.\n\n### **Part 2 — An AI agent runs every creator relationship**\n\nEvery creator on TryNearby is managed over iMessage by an agent we built. Onboarding, match invites, booking, reminders, reschedules, questions. It handles over 10,000 texts a month.\n\nIt also learns as it goes. Every conversation teaches it more about that creator: what they like to eat, where they'll actually drive, when they're free, what they turn down. That context feeds the matching, so the invites get better the longer someone is on the platform.\n\nFor creators, it's a personal assistant that helps them become better creators while bringing collaborations straight to their iMessage instead of them chasing restaurants.\n\n### **Part 3 — Restaurant owners get their own AI agent to text**\n\nOwners get an assistant in iMessage too. It watches every video, comment, and review, and tells them what's happening: who's visiting this week, how the last video performed, what people are saying. So when a dish sells out the week after a creator posted about it, the owner knows exactly why.\n\nOur team sits behind both agents. Anything either one can't handle gets flagged to us in seconds.\n\n![uploaded image](/media/?type=post&id=110956&key=user_uploads/3490413/5ff964f8-9013-42d7-9617-29d6ccb2164a)\n\n## **Traction**\n\n10 months in:\n\n* 120+ paying restaurants across Southern California\n* 37% month-over-month growth with over 90% monthly restaurant retention since November\n* 2,500+ videos posted by local creators → \\~35M views, almost entirely from people within a few miles\n\nWhat that turned into for the restaurants:\n\n* A restaurant in San Diego nearly doubled revenue in a single month\n* Another had a whole new demographic of customers drive a 50% increase in sales\n* Dishes featured in creator videos hit record sales the following week\n\nEvery single owner tells us the same thing: **people are walking in and saying they saw them on TikTok.**\n\n![uploaded image](/media/?type=post&id=110956&key=user_uploads/3490413/3301e743-2fef-43c4-9138-aaa5631d1eb8)\n\n## **Team**\n\n**Yousef** co-founded FaZe Clan after starting as a 14-year-old YouTuber in 2010. He spent 15 years in the creator economy, built an audience of 10M+ followers and over a billion views, and helped take FaZe public on the Nasdaq at a $1B valuation.\n\n**Ahmad** co-founded one of the first social media clipping companies, where he built the first and largest infrastructure for short-form clipping, tracking 10B+ views across tens of thousands of videos for some of the biggest creators on the internet. UC Berkeley CS, previously a senior engineer at WeightWatchers.\n\n**Obaida** co-founded the clipping company alongside Ahmad. He was an early employee at Replit, previously Goldman Sachs, and grew a consumer app from $360K to $2M ARR in 18 months. Master's in applied economics from Johns Hopkins.\n\n![uploaded image](/media/?type=post&id=110956&key=user_uploads/3490413/031126b7-4cbd-49dc-989f-ef79381b8bd9)\n\nIf you own a restaurant, know an owner who wants more foot traffic, are a local creator, or are just interested in TryNearby, we'd love to talk → 📧 [hello@trynearby.com](https://mailto:hello@trynearby.com)", "url": "https://www.ycombinator.com/launches/Src-trynearby-every-local-business-needs-to-be-discoverable-that-150b-market-is-moving-toward-local-creators", "votes": 50, "created_at": "2026-08-17T18:03:07.676Z" } ] }, { "id": 32816, "name": "Financial Datasets", "slug": "financial-datasets", "former_names": [ "Financial Datasets, Inc." ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/4b4bdae68fdd8f9d994fab11fca6b30709eb7e7c.png", "website": "https://financialdatasets.ai/", "all_locations": "New York City, NY, USA", "long_description": "Financial Datasets is an applied AI research lab building search and data infrastructure for agents, starting with financial markets.\r\n\r\nWe turn filings, company disclosures, and market data into structured, searchable information - helping agents find relevant facts, answer financial questions, and monitor markets in real time.\r\n\r\nFor 40 years, market data was built for people reading screens. We're rebuilding it for agents that research every company, all day, across the world.", "one_liner": "Connect your agents to the stock market ", "team_size": 1, "industry": "B2B", "subindustry": "B2B", "launched_at": 1786382003, "tags": [ "Fintech", "Investing", "AI", "Stocks" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/financial-datasets", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/financial-datasets.json", "year_founded": 2024, "city": "New York City", "country": "US", "linkedin_url": "https://www.linkedin.com/company/financial-datasets", "x_url": "https://x.com/findatasets", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/4b4bdae68fdd8f9d994fab11fca6b30709eb7e7c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 2001420, "full_name": "Virat Singh", "title": "Founder/CEO", "founder_bio": "Founder at Financial Datasets. 90,000+ stars on financial agent projects on GitHub. Previously tech lead at Faire, building search recommendations. Earlier software engineer at Airbnb (Payments, Search) and Acorns. Started career in buy-side M&A", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/singhvirat/", "x_url": "http://twitter.com/virattt", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9302e20b46784f462363671342e35fdaf1f8e9e3.jpg" } ], "jobs": [], "news": [ { "title": "The 10 Best Financial APIs for AI Coding Tools (2026)", "url": "https://medium.datadriveninvestor.com/the-10-best-financial-apis-for-ai-coding-tools-2026-21b4caeec56e", "date": "Jun 18, 2026" } ], "launches": [ { "id": 110463, "title": "Financial Datasets: Stock market data for AI agents", "tagline": "The best way to connect your agent to the stock market", "body": "Hey all,\n\nVirat here, founder of [Financial Datasets](https://financialdatasets.ai/) (S26).\n\nWe make it easy for your agents to access live market data for 27,000+ global stocks through a clean and predictable API.\n\nToday, hedge funds and companies building finance agents use [Financial Datasets](https://financialdatasets.ai/) as their market data layer. We are trusted by 1,000+ customers, serving millions of requests daily.\n\n\n\n**The problem**\n\nBloomberg, FactSet, and S&P spent 30 years building market data for humans staring at terminals. The entire stack assumes an analyst in the loop to interpret and verify the numbers.\n\nAgents can't use that stack. Instead, they need machine-readable schemas, natural language search, and a citation back to the source document. None of the incumbent infrastructure was built for the AI era.\n\n**What we built**\n\nThe only market API your agent needs:\n\n* A single place for your agent to gather all of its market data: fundamentals, earnings, KPIs, SEC filings, ownership, pre-IPO financials, prices, and more.\n* Live in seconds, not minutes. Financial data is parsed, structured, and queryable seconds after companies publish them.\n* Benchmarked against ground truth at 99.99% precision. Anything below that bar never reaches the API, so your agent never sees bad data.\n* Delivered over REST, an MCP server, and webhooks. Works out of the box with Claude, ChatGPT, Hermes, and OpenClaw.\n\nYou can ship our data inside your own product. Redistribution included, no per-seat pricing, no anti-AI usage restrictions.\n\n**The team**\n\nI'm Virat, founder of [Financial Datasets](https://financialdatasets.ai/). I have spent 14 years working as a software engineer at tech companies like Acorns, Airbnb, and Faire.\n\nI started Financial Datasets after getting frustrated by how difficult it was to give my agents high quality, real-time stock market data.\n\nI have been building finance agents since 2022 and posting about it daily on [X](https://x.com/virattt). My two open source projects, ai-hedge-fund (60K stars) and Dexter (27K stars), are used by professional and retail investors today.\n\n**Our ask**\n\nIf you’re building agents that touch the market, I’d love to chat. You can also try [Financial Datasets](https://financialdatasets.ai/) right now - it takes 30 seconds to set up and pull data from our API.", "url": "https://www.ycombinator.com/launches/Sjf-financial-datasets-stock-market-data-for-ai-agents", "votes": 6, "created_at": "2026-08-19T20:00:00.465Z" } ] }, { "id": 32848, "name": "Luca IQ", "slug": "luca-iq", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/05718bc482e46fde4208c29e7eef28f2f4b686e7.png", "website": "https://lucaiq.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Luca IQ is an API for deterministic tax calculation and an IRS Authorized e-filing engine - the foundation for accounting firms and fintechs building on tax.", "one_liner": "API-first tax engine for CPA firms and developers", "team_size": 5, "industry": "B2B", "subindustry": "B2B -> Finance and Accounting", "launched_at": 1785364349, "tags": [ "B2B", "API", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Finance and Accounting" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/luca-iq", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/luca-iq.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/lucaiq", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/05718bc482e46fde4208c29e7eef28f2f4b686e7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/b7319b807eb27608540596ccef27337c2998ea22.png" ], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 2754880, "full_name": "Angelo Policicchio", "title": "Founder/CEO", "founder_bio": "Founder at Luca IQ. CPA. Previously Equity Research Associate at Credit Suisse, Forensic Accounting at the FBI, Public Accountant at EY. MBA from Harvard Business School.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/angelo-policicchio", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/56e0986c5f092107d5de09990091ea3d9a92d763.jpg" }, { "user_id": 3101824, "full_name": "Akash Sadashivapeth", "title": "Founder/CTO", "founder_bio": "Founder at Luca IQ. Previously Head of Data Science at Welligence Energy Analytics.\n\nStudied Data Science at Penn and an MBA from Harvard Business School.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/akash-sadashivapeth/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/19e9446ccd10b06ab85431a7e79b07c759587087.jpg" } ], "jobs": [ { "id": 108815, "title": "Founding Tax Lead (CPA or EA Required)", "url": "https://www.ycombinator.com/companies/luca-iq/jobs/VQUCBls-founding-tax-lead-cpa-or-ea-required", "location": "Chicago, IL, US / Remote (US)", "type": "Full-time", "role": "Product", "role_type": null, "salary_range": "$100K - $200K", "equity_range": "0.25% - 1.50%", "experience": "6+ years", "visa": "US citizen/visa only", "skills": [] }, { "id": 110614, "title": "AI Engineer", "url": "https://www.ycombinator.com/companies/luca-iq/jobs/U8IX0ke-ai-engineer", "location": "Chicago, IL, US", "type": "Full-time", "role": "Engineering", "role_type": "Machine learning", "salary_range": "$150K - $200K", "equity_range": "", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [ "Python", "Machine learning", "LLMs", "Fine-tuning", "Evals", "AI Agents" ] } ], "news": [ { "title": "Luca IQ receives IRS approval for federal 1040 e-filing", "url": "https://cfotech.news/story/luca-iq-receives-irs-approval-for-federal-1040-e-filing", "date": "Aug 03, 2026" }, { "title": "Tech news: Luca IQ launches with e-file approved tax engine for CPA firms | Accounting Today", "url": "https://www.accountingtoday.com/list/tech-news-luca-iq-launches-with-e-file-approved-tax-engine-for-cpa-firms", "date": "Jul 31, 2026" } ], "launches": [ { "id": 108364, "title": "Luca IQ - An API for U.S. tax calculation and e-filing", "tagline": "An open API, IRS-authorized e-file tax engine for CPA firms and Fintechs", "body": "# \\*\\*\\*\\*\n\n**TL;DR: We're building a modern tax engine for accountants and fintechs building on tax: a deterministic tax calculation and e-filing core behind a single API, built on the IRS's official e-file infrastructure (MeF).**\n\n## A tax return is math. It should be a function call.\n\nWe're Angelo and Akash, and we built Luca IQ, a new tax engine for accounting firms. One set of inputs in, a completed, filed return out in minutes. Not bolted onto decades out desktop software.\n\nFor CPA firms, that means more returns per preparer, less spend on labor, and a clean path off 30-year-old desktop software. For the PE groups acquiring and scaling firms, it means real operating leverage on the single most labor-intensive part of the business.\n\n\n\n**The problem.** \\~95% of US firms file their returns through just four companies, and all of them were built decades ago, for the desktop, and are closed. Firms are stuck running tax season on software that can't scale with them. Every modern AI tax tool just pushes data into that same closed software it can't control. Everyone rebuilt the interface.\n\n**What we built.** A new one, and we're opening it. A key principle: AI never does the tax math. AI and human judgment set the inputs, and a deterministic engine computes the return, with a full audit trail on every number. It's either 100% correct or it's wrong, and only a deterministic engine can promise the former. For firms, that means throughput without giving up accuracy or defensibility.\n\nOur core leverage: we're an **IRS-authorized e-file provider as a Software Developer, Transmitter, and Intermediate Service Provider**, so Luca IQ doesn't just compute the return, it e-files straight to the IRS. That full path, calculation through filing, is what firms can't get from a closed engine today.\n\n**Why now.** The profession is running out of people. 300,000+ accountants have left since 2019, and \\~70% of CPAs are near retirement. At the same time, private equity is rolling up firms and pushing hard on margin. The work isn't going away. The people are, and the pressure to do more with fewer preparers has never been higher.\n\n**Who we are.** Angelo (CEO) is a CPA who worked in forensic accounting with the FBI, Wall Street, EY and spent years working at tax firms using the legacy softwares we are now replacing. Akash (CTO) was employee #1 at a fintech data-infrastructure startup and leads engineering.\n\n**What we need:**\n\n* **CPA firms who dread tax season**, We'd love 20 minutes to show you what Luca IQ does for throughput and margin. This is who we build for.\n* **PE firms rolling up accounting practices.** If tax prep labor is your biggest cost line, let's talk.\n* **Developers and fintechs** who need 1040 calculations or e-filing. The engine is one API call, and you can build on it early.\n\nReach us at [founders@lucaiq.com](mailto:founders@lucaiq.com).", "url": "https://www.ycombinator.com/launches/SBo-luca-iq-an-api-for-u-s-tax-calculation-and-e-filing", "votes": 4, "created_at": "2026-07-31T14:01:23.291Z" } ] }, { "id": 32913, "name": "Pacific", "slug": "pacific", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/196624bfab919b6dd989cb601af1771954cb562e.png", "website": "https://pacific.space", "all_locations": "San Francisco, CA, USA", "long_description": "We mass-produce modular data centers that deploy on stalled infrastructure projects losing billions per week.\r\n\r\nDelayed infrastructure projects hemorrhage $1B a week in the US alone - Pacific mass-produces modular data centers that deploy on those sites in days and turn their unused power into revenue.\r\n\r\nData centers are stuck for years waiting for power, while delayed construction sites across the world sit on unused interconnects, losing billions of dollars per week in interest on trillions of dollars of debt. Our pods deploy on those sites and generate revenue for them in days. They use diamond wafers in the cooling system so pods run anywhere on earth, even a desert. \r\n\r\nWe'll use the same mechanisms we're building today to deploy compute across the galaxy.\r\n", "one_liner": "Mass-producing and financing micro data center incredibly quickly.", "team_size": 6, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1786300632, "tags": [ "Fintech", "Robotics", "Space Exploration", "Manufacturing", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/pacific", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/pacific.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/pacific-compute", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/196624bfab919b6dd989cb601af1771954cb562e.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3218344, "full_name": "Harper Dammann Smith", "title": "Founder/CEO", "founder_bio": "Founder and CEO of Pacific - Grew my drone startup to $1M in revenue at 18; left school to shoot expeditions for Red Bull on 7,000m peaks. Studied aerospace engineering; dropped out after selling my first company. Now mass producing modular data centers that deploy in days turning unused power into Ai factories.\n\nI'll go to space one day.\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/harper-dammann-smith/", "x_url": "https://x.com/harperdammann?s=21", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3e20832d513b9aced0007904f88c4e05ba173221.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109996, "title": "Pacific: Mass-producing and deploying micro data centers for Earth and Space", "tagline": "Pacific mass produces micro data centers for Earth and space, faster than anyone else.", "body": "**Building the toughest super computers for earth and space**\\\n\\\n**TL;DR:** Data centers are stuck for years waiting for power, while delayed construction sites across the world sit on unused interconnects, losing $1B per week in the US alone, in interest on trillions of dollars of debt. Our pods deploy on those sites and generate revenue for them in days. They use diamond wafers in the cooling system so pods run anywhere on earth, even a desert.\n\n\n\n**The deal**\n\nIf you need compute (GPU hours, tokens, capacity for a training run) send us a note. We sell short and long-term contracts below market rate.\n\nIf your data has to stay on-prem (**defense, CUI, CMMC)**, we sell managed, modular pods directly and are already working with large aerospace companies.\n\n**We're building factories**\n\nSpaceX didn't just build rockets, they built rocket factories. We do the same and our first factory stands up over the next few months, potentially funded by the state of Missouri. We vertically integrate the critical components that currently carry multi-year lead times: transformers, switchgear and power distribution, and by taking control of our own supply chain we ensure we can build pods fast enough.\n\n**Financed like aircraft**\n\nOur pods are financed like aircraft - debt at **\\~5-6%, not 11-15%**. A data center is a risky construction loan: money stuck in a half-built site for years, and if it fails the lender holds a building and a pile of GPUs priced by a brutally volatile market - the banks are freaking out. Our pods leave the factory fully functional, earns the moment it's connected to power, and can be repossessed and redeployed: the profile of equipment finance, far faster and cheaper.\n\n**Traction**\n\nToday we have $95M in LOI’s for our hosted compute and a further $27M for direct sales of our pods. Our first two clusters go live in Missouri this month, followed by the first modular factory.\n\n**Our asks**\n\n* **Compute buyers and offtake.** We'd love to speak to anyone buying compute, or setting up on-prem deployments, or needing CMMC compliance. Introductions to procurement teams at OpenAI, Anthropic, or any startup constrained by access to cheap compute would be incredible.\n* **Hiring.** Data center engineers, electrical engineers, any data center nerds you know.\n* **Financiers.** These projects run on capital from infra investors and tech lenders. If you've worked in that world, we'd love to chat.\n* **Construction sites.** If you sell to construction sites, work in construction, mining, or fab production, or own a site yourself - we will put a money machine on it. Let's chat.\n\n**Join us as we build supercomputers for the galaxy!**\n\n![uploaded image](/media/?type=post&id=109996&key=user_uploads/3218344/e8f50721-c56b-4585-9abf-7d11cf11d5c3)\n\n", "url": "https://www.ycombinator.com/launches/Sc8-pacific-mass-producing-and-deploying-micro-data-centers-for-earth-and-space", "votes": 10, "created_at": "2026-08-12T17:30:05.717Z" } ] }, { "id": 32914, "name": "Last Accounting Company", "slug": "last-accounting-company", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/cf21d2add4951b9271177840cd785b32579fdb0f.png", "website": "https://lastaccountingcompany.com/", "all_locations": "Helsinki, Finland", "long_description": "Last Accounting Company (LAC) is a full-stack AI accounting firm. We sell the same service (bookkeeping, tax filings, etc.) as a traditional accounting firm, but underneath agents handle everything and a certified accountant signs off. Our service is faster, cheaper, and fully integrated, allowing our clients to focus on what matters, not accounting admin.\r\n\r\nLAC replaces both outsourced accountants and legacy accounting software for startups and SMBs. Our system is fully proprietary, from the general ledger to the accounting agents. This allows us to create powerful self-reinforcing loops and ultimately provide a state-of-the-art accounting service.", "one_liner": "AI-native accounting firm.", "team_size": 4, "industry": "B2B", "subindustry": "B2B -> Finance and Accounting", "launched_at": 1784575015, "tags": [ "Fintech", "Finance", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Finance and Accounting" ], "regions": [ "Finland", "Europe" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/last-accounting-company", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/last-accounting-company.json", "year_founded": 2026, "city": "Helsinki", "country": "FI", "linkedin_url": "https://www.linkedin.com/company/lastaccountingcompany", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/cf21d2add4951b9271177840cd785b32579fdb0f.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/d8ee1ef9812e159b693b8ea9e2dca2f73efbf278.png" ], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 3677074, "full_name": "Sami Laakkonen", "title": "Founder", "founder_bio": "CEO & Co-Founder at Last Accounting Company. Built petabyte-scale data infra for clients incl. Google and Bloomberg. 2x Nordplus International Science Olympiad in Physics. Ex-fighter pilot training. CS dropout.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/sami-laakkonen/", "x_url": "https://x.com/samilaa", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c78fe557404b8f7ef13811329fbe25bb5f686606.jpg" }, { "user_id": 2422668, "full_name": "Riki Karjalainen", "title": "Founder", "founder_bio": "Co-Founder at Last Accounting Company. Previously built petabyte-scale data infrastructure at Token Terminal. Applied Mathematics at Aalto University.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/riki-karjalainen-22a612189", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e6694a4980bf011635833fac045307bbc3efbf34.jpg" }, { "user_id": 3682484, "full_name": "Veikko Pere", "title": "Founder", "founder_bio": "Co-Founder at Last Accounting Company. Previously at BCG and private equity at Intera Partners with IB and law internships. M.Sc. Finance and M.Sc. Laws (ongoing).", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/veikkopere/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/60197292925f48ab44361841ce811633735344c7.jpg" }, { "user_id": 3657494, "full_name": "Pietari Virkkunen", "title": "Founder", "founder_bio": "Co-Founder at Last Accounting Company. Prev. Tech/PE/VC. M.Sc. Industrial Engineering and Management at Aalto University.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/pietarivirkkunen/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/530764dc365306895ac5eeed832500b2759e5ef8.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111616, "title": "The Last Accounting Company you’ll ever need", "tagline": "Agent-native accounting firm that replaces the legacy software and the outsourced service.", "body": "Hey YC 👋\n\nWe’re[ Riki](https://www.linkedin.com/in/riki-karjalainen-22a612189),[ Sami](https://www.linkedin.com/in/sami-laakkonen/),[ Veikko](https://www.linkedin.com/in/veikkopere/), and[ Pietari](https://www.linkedin.com/in/pietarivirkkunen/), the founders of[ Last Accounting Company](https://lastaccountingcompany.com/).\n\n**TL;DR**\n\n**LAC is the agent-native accounting firm. \\\nAI agents do the work and our (human) accountants validate all outputs.**\n\nWe replace both the traditional accounting firm and the legacy accounting software with one end-to-end service. Founders get current books, correct filings, and an always up-to-date general ledger, not another dashboard to operate.\n\n[https://drive.google.com/file/d/100N7xvBREGMXpbgnYOUNAiFqht0Gc3EA](https://drive.google.com/file/d/100N7xvBREGMXpbgnYOUNAiFqht0Gc3EA/view?usp=sharing)\n\n**🗃️ The problem**\n\nFounders didn’t start companies to learn accounting. Yet disconnected systems turn them into the human link between their accountant, bank, payroll, invoices, and internal tools.\n\nThey find and forward data, wait for questions, track down missing context, and repeat the cycle. By the time the books are closed, the numbers are often too old to guide day-to-day decisions.\n\n**🛠 Our solution**\n\nLAC connects directly to your bank, payroll, and internal systems. Our agents:\n\n* Keep the general ledger continuously up to date\n* Handle bookkeeping, invoices, payroll, and receipts\n* Prepare filings according to local requirements\n* Surface missing information and potential issues early\n\nOur expert accountants validate the work before anything is filed.\n\nLAC also connects financial data with the tools teams already use, including ERP, CRM, and product analytics. Through our Slack agent or MCP, founders can ask questions like:\n\n* What are our burn and runway?\n* How’s our cash flow looking for Q3? Can we afford a new hire?\n* Where did spending change this quarter?\n\n**Your books stop being only a filing obligation. They become the ground truth behind every decision.**\n\nUnlike accounting tools built for one task at a time, LAC is one system—from the general ledger to the accounting agents—and we operate the service ourselves.\n\nCustomers simply get their accounting done.\n\n**👥 The team**\n\nWe’re three technical builders and one finance-and-law brain. Riki and Sami previously built petabyte-scale financial data infrastructure at Token Terminal. Veikko brings experience from BCG, private equity, finance, and law. Pietari brings engineering and sales experience across tech, VC, and PE.\n\n**🙏 Our ask**\n\nRun a startup or SMB in Europe or the US, or know someone who does?\\\nWe’d love to show you what LAC can do! A one-line intro is enough and we’ll take it from there.\n\nEmail us at: [hello@lastaccountingcompany.com](mailto:hello@lastaccountingcompany.com)\\\nSign up or read more at:[ ](https://lastaccountingcompany.com/)[lastaccountingcompany.com](http://lastaccountingcompany.com)", "url": "https://www.ycombinator.com/launches/T2G-the-last-accounting-company-you-ll-ever-need", "votes": 70, "created_at": "2026-08-21T17:08:00.094Z" } ] }, { "id": 32947, "name": "Executor", "slug": "executor", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/948ce8964e024f509e8d2c3db33ab94f3acb5b17.png", "website": "https://executor.sh", "all_locations": "San Francisco, CA, USA", "long_description": "Executor.sh is an MCP gateway that allows you to sign in once to all of your integrations and use them with any client.\r\n\r\nWe enable you to connect your agents to anything (MCP, GraphQL, OpenAPI) so regardless of what you're trying to connect agents to you're able to.\r\n\r\nIf you're trying to scale agents to access all of your production services, data, etc, you need Executor as a unified management layer.", "one_liner": "The open source tool gateway", "team_size": 1, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1785345907, "tags": [ "AIOps", "SaaS", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/executor", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/executor.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "https://x.com/executorsh", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/948ce8964e024f509e8d2c3db33ab94f3acb5b17.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1133501, "full_name": "Rhys Sullivan", "title": "Founder", "founder_bio": "Founder at Executor, enabling your to connect every agent to anything. Previously Software Engineer at Vercel, at Microsoft, Raven Software, and Epic Games.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rhyssullivan/", "x_url": "https://x.com/RhysSullivan", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/91ca30bf14b18325ab55dfdc29d5723fab969094.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108331, "title": "Executor — One place to connect your agents to everything", "tagline": "Connect your tools once and every agent has them, with permissions you control", "body": "**TL;DR:** Executor is your tool gateway. Connect every service and every account once, set per-tool permissions, and any MCP-compatible agent (Claude Code, Codex, Cursor, whatever ships next) reaches all of it through one endpoint. Try it at [executor.sh](https://executor.sh).\n\nhttps://youtu.be/NWYFpDW02Dw\n\n## The problem\n\nYou don't have one agent anymore. You have Claude Code, a ChatGPT account, Cursor, maybe a box in the corner running jobs overnight. Three problems follow:\n\n1. **Every agent wants its own OAuth into everything.** You connect your calendar to the tenth harness in a slightly different way, tokens expire, connectors deprecate, and each agent is missing something when you need it.\n2. **Most connectors stop at one account per service.** One Google login means your agent sees your work calendar or your personal one, never both. Your flights land in one inbox, your meetings in the other, and no agent can reason over your actual day.\n3. **Permissions are an afterthought.** Once an agent holds a token, it can do anything that token allows. Giving an agent your email shouldn't mean trusting it to never hit delete.\n\n## What Executor does\n\nConnect once in Executor, and sever all the one-off connections.\n\nEach integration gets credentials plus a per-tool policy: **always allowed**, **needs approval**, or **blocked**. Point any MCP-compatible agent at Executor and it has everything.\n\n- **All your accounts, not just one.** Accounts are connections: work and personal Gmail, two calendars, a staging and a production Stripe, side by side. Every agent sees all of them.\n- **One tool shape.** MCP, OpenAPI, and GraphQL integrations all normalize into a consistent name + schema. 1,600+ tools collapse into a single ~1k-token surface instead of ~280k tokens of tool definitions.\n- **Safe by default.** Executor preserves operation semantics (GET vs DELETE, destructive hints, GraphQL mutations), so agents auto-run safe operations and ask before risky ones.\n- **Sandboxed execution.** Tool calls run in an isolated JavaScript sandbox; secrets are injected host-side, so the model never sees a raw token.\n- **Team-ready.** Per-user and shared credentials. Onboard the team once, not per agent per person.\n\nBecause enforcement lives in Executor rather than in any agent framework, you can adopt whatever new agent ships next week without re-thinking data access. You keep everything first-party connectors offer and gain what they don't: multiple accounts per service, per-tool policies, and one place to see and revoke access.\n\n## How to use it\n\n- **Local:** open-source (MIT) desktop app and CLI, everything stays on your machine\n- **Self-hosted:** a Dockerfile that lets your team own all its data\n- **Executor Cloud:** hosted, nothing to manage\n\n## Traction\n\n- 2,800+ GitHub stars, ~14k npm downloads/week, ~3,000 users on the local version\n\n## Background\n\nSolo technical founder, previously at Vercel. Before Executor I built [answeroverflow.com](https://www.answeroverflow.com) and grew it to 1.5M monthly users as a solo dev, bootstrapped.\n\n## Asks\n\n1. **Try it and tell me where it breaks.** [executor.sh](https://executor.sh) has a \"Copy prompt\" setup you can paste into your agent to get started.\n2. **Intros:** if your team (or one you know) is drowning in per-agent integration setup, I'd love an intro: rhys@executor.sh\n3. **Star the repo:** [github.com/UsefulSoftwareCo/executor](https://github.com/UsefulSoftwareCo/executor)\n4. **Want a walkthrough?** Grab a time: [cal.com/rhys-sullivan/executor-chat](https://cal.com/rhys-sullivan/executor-chat)\n", "url": "https://www.ycombinator.com/launches/SBH-executor-one-place-to-connect-your-agents-to-everything", "votes": 9, "created_at": "2026-08-04T15:41:42.736Z" } ] }, { "id": 32948, "name": "Waybill", "slug": "waybill", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/b91adedb53afabc881281324143485f1dc3c6674.png", "website": "https://www.waybill.to/", "all_locations": "Bengaluru, KA, India", "long_description": "Hardware teams burn weeks on procurement. An engineer needs a part. Someone emails five distributors, chases quotes, gets an approval, wires the money, tracks a shipment across three carriers, clears customs, and then nobody knows where the box is or what's actually on the shelf.\r\n\r\nWaybill runs that entire loop. Send us a part, a spec, or a BOM. We source it, pull quotes, route approvals, pay the supplier, handle freight and customs, track it to your dock, and keep your stock count updated. \r\n\r\nTake your supply chain for granted.\r\n\r\nBuilt by a team that's seen it at LAT Aerospace and WareIQ (YC S20).", "one_liner": "Agentic procurement & inventory for deep tech teams.", "team_size": 3, "industry": "B2B", "subindustry": "B2B -> Supply Chain and Logistics", "launched_at": 1786690936, "tags": [ "Logistics", "Procurement", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Supply Chain and Logistics" ], "regions": [ "India", "South Asia" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/waybill", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/waybill.json", "year_founded": 2026, "city": "Bengaluru", "country": "IN", "linkedin_url": "https://www.linkedin.com/company/waybill-inc", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/b91adedb53afabc881281324143485f1dc3c6674.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1071184, "full_name": "Rishi Laddha", "title": "Founder/CEO", "founder_bio": "Founder at Waybill. Previously managed operations, supply chain, and government relations at LAT Aerospace (founded by Deepinder Goyal of Zomato). Built and scaled AI trade assistants at INDmoney. BTech in CS & AI from Plaksha University.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rishiladdha", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/50bddf9b35eb2d7b15a92974b396c1a57b8d634f.jpg" }, { "user_id": 1476118, "full_name": "Tushar Goyal", "title": "Founder", "founder_bio": "Founder at Waybill (S26). Previously built AI and software at WareIQ (S20). Computer Science, Plaksha University.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/thetushargoyal/", "x_url": "https://x.com/thetushargoyal", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fd749c59593d9ac4681a39e6a981f1d0f291cd1b.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111330, "title": "Waybill - Procurement autopilot for hardware teams.", "tagline": "Take your supply chain for granted.", "body": "Hi YC 👋 we're Rishi and Tushar, and we're building Waybill ().\n\n**TL;DR:** Waybill runs procurement for hardware and deep-tech companies, end to end. An engineer tells us what part they need. Our agents and humans source it anywhere in the world, get quotes, handle payment, shipping, customs, last mile, log it into inventory and match the invoices. One system of record from request to stock. We don’t stop at the ‘order placed’. We make sure it reaches your shelf. A few crazy things Waybill was asked to source & order -\n\n1. A lifetime supply of an obsolete RF component\n2. Cessna aircraft. Yes, real airplanes\n3. Industrial Robots\n\nLaunch video: \n\n**Backstory**\n\nI (Rishi) ran operations at LAT Aerospace (Deepinder Goyal’s aerospace startup) as part of their founding team. On paper, that meant \"sourcing.\" In practice, it meant finding jet engines, flight controllers, high-density batteries, and satellite radios for a very ambitious aerospace program, then moving machinery and merchandise worth millions across borders, and finding and operating real estate you can actually fly aircraft from. I also handled government liaisoning, which is a polite way of saying I convinced the world's harshest bureaucracy to approve things it didn't have a form for, like taking off and landing an aircraft in under 50 meters.\n\nTushar has been working the other end of this problem his whole career. At WareIQ (YC S20), he built the autonomous inventory system behind 400+ Indian D2C brands - forecasting demand and making buying decisions across thousands of volatile SKUs without blowing up working capital.\n\nSo between us: I've spent my career buying hardware the hard way, and Tushar's spent his teaching machines to manage inventory. Waybill is both jobs in one system.\n\n**The problem**\n\nHardware engineers spend a big chunk of their week buying parts. They can't hand it off to operations folks, because only the engineer who specified the part knows which alternates are acceptable. And finding parts is genuinely hard: stock is scattered across distributor catalogs, resellers, and Alibaba listings, and packaging suffixes, obsolete part numbers, and regional inventory make the same part look like six different parts. It's a search, not a lookup.\n\nAll of this runs on email, WhatsApp, and spreadsheets. No system of record, and context dilutes at every handoff. Parts land late, arrive undocumented, or get bought twice. I know because at LAT, _I was the system of record_. It didn't scale.\n\n**What we built**\n\nAn engineer tells Waybill what they need. Our agents find every authorized distributor and manufacturer rep stocking that part anywhere in the world, request quotes and lead times, and rank the options by what the customer actually cares about. A defence company only wants suppliers from allied nations. A medical device company needs provenance. Everyone else wants the right balance of cost, lead time, and complexity.\n\nIf no distributor has it, our network in India, China, and Southeast Asia finds it. If a supplier can't ship to the customer's country, we pick up the box and ship it ourselves.\n\nOnce the order is placed, our agents prepare the customs paperwork, a licensed broker files it, parts are received into inventory automatically, and the invoice is automatically matched against the PO and import documents. The next time someone asks for that part, the agent says it's already on the shelf.\n\nAnd when a customer moves from R&D to production, the same system scales with them. Demand comes off the BOM and build schedule - reorder points, second sources, volume pricing - and we already hold every order, price, and lead time they've ever had.\n\nMost procurement software ends at an approved PO. We end at a physical part on a shelf, customs-cleared and invoice-matched.\n\nRippling made the employee the first-class object and built every service around it. We're doing the same for the part.\n\nHere are a few crazy things engineering teams asked Waybill to procure in the two weeks that we’ve been live -\n\n1. A lifetime purchase of an obsolete RF component\n2. Cessna aircraft. Yes, real airplanes.\n3. Industrial Robots\n\n**Where this goes**\n\nA hardware company should be able to ask us for anything physical and get it. Today that's parts. Next it's PCB fab and assembly, CNC machining, 3D printing, tooling, trucking, warehousing, and allocation on ships and planes. The interface never changes: say what you need, and it arrives.\n\n**Asks**\n\n1. Hardware and deep-tech founders - we want to run your procurement. Comment, DM, or email me at [rishi@waybill.to](mailto:rishi@waybill.to). We’re company size agnostic.\n2. Intros to procurement leads at hardware enterprises. If you know one, we'd love a connect.\n\n**Ask for it and it shows up. Take your supply chain for granted.**", "url": "https://www.ycombinator.com/launches/Sxe-waybill-procurement-autopilot-for-hardware-teams", "votes": 13, "created_at": "2026-08-20T03:30:00.319Z" } ] }, { "id": 32979, "name": "RonanRx Inc.", "slug": "ronanrx-inc", "former_names": [ "Ronan Rx Inc." ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ce1dcec626f9c7fdc1ea66f4d7850c7a32015773.png", "website": "https://ronanrx.com/", "all_locations": "San Francisco, CA, USA", "long_description": "RonanRX is a full-stack pharmaceutical company, starting with personalized peptides and GLP-1s.\r\n\r\nIn our first 10 weeks, we went from having no pharmacy to being licensed in 33 states. Since launch, we’ve grown 50% week over week.\r\n\r\nWe manufacture the medications ourselves, sell directly to patients, and eliminate the layers of middlemen that make prescription drugs more expensive.\r\n\r\nWhat makes this possible is AI.\r\n\r\nTraditional pharmaceutical companies are built to manufacture one drug for millions of people. RonanRX is purpose-built to personalize treatment for each patient: different biology, different response, different dose.\r\n\r\nPatients come directly to RonanRX or through doctors using our platform. Every patient completes our medical intake, meets with a licensed doctor through our telehealth platform, and receives a personalized prescription when appropriate. From there, we manufacture the medication ourselves and deliver it directly to the patient.\r\n\r\nLabs, wearables, medical history, treatment response, and other health data feed back into the platform so doctors can continuously refine treatment over time.\r\n\r\nBefore AI, doing this for every patient at scale was impossible.\r\n\r\nWe didn't invent the demand we are seeing. 80 Million people are on a GLP-1 and millions more are using Peptides. We're building a better way to manufacture, personalize, prescribe, and deliver them.\r\n\r\nOur founders have built and exited multiple companies, raised more than $100M, built a $50M U.S. medical manufacturing business, and spent years developing peptide therapeutics at companies including Eli Lilly and AbbVie.\r\n\r\nWe’re just getting started with peptides and GLP-1s. Next comes hormone therapy and the rest of personalized medicine.", "one_liner": "Full Stack Pharmaceutical Company for Personalized Peptides and GLP-1s", "team_size": 5, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1784908687, "tags": [ "Health Tech", "Telehealth", "Therapeutics" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/ronanrx-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/ronanrx-inc.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/ronanrx/", "x_url": "https://x.com/Ronan_RX", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ce1dcec626f9c7fdc1ea66f4d7850c7a32015773.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 26799, "full_name": "Lloyd Armbrust", "title": "Founder/CEO", "founder_bio": "2x YC founder. 3 Exits. 6 Kids. Built a $50M manufacturing business. Now building RonanRX.\n\nWe are the world's first full-stack pharmaceutical company for personalized peptides & GLP-1s. We manufacture and ship direct to patients, using AI to personalize treatment. This was impossible before AI. \n\nLicensed in 33 states; growing 50% WoW.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/lloydarmbrust/", "x_url": "https://x.com/larmbrust", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b0cd6642a974754f16fc7079423c9310a14a4cc6.jpg" }, { "user_id": 3672671, "full_name": "Emily Colman", "title": "Founder", "founder_bio": "Co-Founder & COO at RonanRx, I deliver personalized, doctor-reviewed peptide and GLP-1 care that starts with a text. \n\nI am go-to-market operator with over a decade in healthcare sales, most recently at Eli Lilly, I support weight management solutions for hundreds of physicians and millions of patients.\n\nI have 2 children: Quinn (4) and Leo (1) based Austin, TX.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/emcolman?utm_source=share_via&utm_content=profile&utm_medium=member_ios", "x_url": "https://x.com/emilyrcolman?s=11&t=ctOmnMNv7bFv0OBgXfv_0g", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/59f45cc4f2e90d9789b60f843b64b980ca729f95.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107336, "title": "RonanRX: We use AI to personalize peptides to your biology", "tagline": "RonanRX is a full-stack AI pharmacy specifically for peptides", "body": "**Peptides are like Claude Code for your body. Your biology is the prompt.**\n\n**TL;DR:** RonanRx personalizes GLP-1s like Ozempic and other peptides to _your_ biology, because you are not average. Connect your lab results, wearable data, and medical history once. Our HIPPA compliant Al turns them into a plan which a licensed doctor prescribes, adjusts every cycle, and ships to your door. Start with a text at [**RonanRx.com**](http://RonanRx.com).\n\n### **Big Pharma Makes 1 Drug for Millions.**\n\nPeptides are the most personalizable medicine we have, and everyone has different biology. GLP-1s like Ozempic help you lose weight. Others help you recover faster, sleep better, live longer. Telemedicine often prescribes them like a vending machine: fill out a quiz, a doctor you’ll never meet rubber-stamps it, and you get the same molecule at the same dose as everyone else. Nobody reads your data and adjusts when your body responds differently than the “average patient.” That’s not medicine, it’s a spreadsheet.\n\n### **How RonanRx works**\n\nConnect your health data once, and we run the complete feedback loop for personalized medicine:\n\n* **Your biology in one place.** We sync your labs, wearables, medical history, and treatment response into a single health profile.\n* **AI turns data into a plan.** It reads how _you_ are responding and proposes the next move: dose, formulation, what to add.\n* **A real doctor decides.** Every prescription is made and adjusted by doctor looking at the same data you are. The AI proposes, the doctor prescribes.\n* **Delivered to your door,** and every outcome improves your next protocol.\n\nInstead of treating the average patient, we personalize your treatment continuously as your body changes.\n\n### **Why us**\n\nBuilding a healthcare company isn’t for tourists. We’re native to the space, and passionate about fixing what’s wrong. Our founders have built and exited multiple companies, raised over $100M, and built a $50M U.S. medical manufacturing business. Lloyd Armbrust is a 2x YC founder, 3 exits, 6 kids, and the oldest founder in S26. Emily Colman spent a decade in healthcare sales, most recently at Eli Lilly, supporting weight-management for hundreds of physicians and millions of patients.\n\n### **Find out what peptides can do for you**\n\n**Start texting us at [RonanRx.com](http://RonanRx.com)** and become the maximum version of yourself.\n\n![uploaded image](/media/?type=post&id=107336&key=user_uploads/26799/336818ca-7507-4c5a-aaaf-9909502fdeed)\n\n_Doctors and clinics: we white-label this entire stack so you can offer personalized peptide care under your own name. Reach us at \\[[**founders@ronanrx.com**](mailto:founders@ronanrx.com)\\]._\n\n— Lloyd Armbrust & Emily Colman", "url": "https://www.ycombinator.com/launches/RvE-ronanrx-we-use-ai-to-personalize-peptides-to-your-biology", "votes": 65, "created_at": "2026-07-24T19:21:00.074Z" } ] }, { "id": 32981, "name": "Marengo", "slug": "marengo", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/60eafb41e958f2dbb87a17e5996bba09e7731f2f.png", "website": "https://marengox.com", "all_locations": "San Francisco, CA, USA", "long_description": "Marengo is the AI-native engineering firm automating engineering design, starting with data centers. We’ve generated $70K in revenue in our first month, with $4M in contracts to deliver.\r\n\r\nMarengo is backed by YC, Exor, Future Planet Capital, undisclosed $2B AUM family office, and other top-tier investors.\r\n\r\nOur team includes the former President & CEO of Saipem America ($3B+ P&L), former President of Exyte's Advanced Technology Facilities GBU ($6B+ P&L) and more.", "one_liner": "AI-Native Engineering Firm designing Data Centers ", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Energy", "launched_at": 1786181708, "tags": [ "Construction", "Design", "Energy", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Energy" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/marengo", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/marengo.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/marengox", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/60eafb41e958f2dbb87a17e5996bba09e7731f2f.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3204597, "full_name": "Emil Ares", "title": "Founder/CEO", "founder_bio": "Founded my first company at 16, making six figures. Ex-Cambridge Physicist working on world's most powerful magnets (56MW) at the 1.4GW MagLab facility. Ex-UK Space Agency.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/emil-ares", "x_url": "https://x.com/emilares", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/73aa5b2d36814a17c6bd8b5b3c2f6e725d9dd15e.jpg" }, { "user_id": 3410964, "full_name": "Gad Marconi", "title": "Founder/CTO", "founder_bio": "Ex-AI software engineer at Nuclear and Energy EPC. Built and deployed on-site data center infrastructure at the 25MW Swiss Flumenthal hydroelectric plant. Ex-TU Delft.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/gad-marconi", "x_url": "https://x.com/gadmarconi", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/70c93b94836254602a740fc2c01e50a852e1acbc.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110380, "title": "Marengo: Accelerated Data Center Design", "tagline": "Marengo designs data centers at double the speed and half the cost with AI.", "body": "Hey everyone 👋! We're [Emil Ares](https://www.linkedin.com/in/emil-ares/) and [Gad Marconi](https://www.linkedin.com/in/gad-marconi/), co-founders of [Marengo](https://www.linkedin.com/company/marengox).\n\n**TL;DR** $6.7T is going towards data center capex up to 2030. It still takes 10-12 months to design data centers. Marengo is _the engineering firm for accelerated data center development._ Site Due Diligence through FEED and permitting design in half the traditional time and half the cost. We’ve developed internal AI design tools that our professional engineers use to accelerate their engineering design work. Customers receive validated outputs they can trust. Already working with some of the largest developers in the world.\n\n---\n\n**Our Launch Video: [https://youtu.be/57EbJ0Buz-U](https://youtu.be/57EbJ0Buz-U)**\n\n**Ask:** Are you a data center developer? Give us a site. Let Marengo show you what accelerated engineering looks like on one of your projects. Click [here](https://marengox.com/).\n\n**The problem\\[s\\]:**\n\nHuge **expenditure on design**: $6.7T of data center capex by 2030 - engineering design fees are a massive 12% of construction cost, meaning $800B in design spend alone.\n\nHuge **time waste**: engineering design takes 12-36 months (feasibility, due diligence, FEED/permitting design, detailed design); it’s done manually, meaning a design firm can only handle 3-4 different concepts per site, leading to suboptimal designs, worse decisions and higher risk.\n\n**The Solution:**\n\nAI alone doesn’t have the engineering, know-how or trust to be effective here. A step change is needed. Marengo is needed.\n\nTo customers, we’re **_the engineering firm for accelerated data center development_.** Internally, we have developed AI-native tooling designed for in-house use by professionals. These tools are:\n\n* **Physics-informed** - grounded in reality\n* **Deterministic** - the same specs will produce the same designs)\n* **Validated** - validated by our professional in-house engineers\n* **Data-driven** - we parallelise concept design generation, producing a virtually unlimited number of variations; we then run simulations to determine the best fit\n* **AI-native** - AI agents act as junior engineers accelerating manual processes so our professional engineers can focus on validation\n\n**How it works:**\n\nA customer gives us their plot of land; we leverage their heritage and requirements to deliver validated designs in **half the time** and **half the cost.**\n\n**Who We Are:**\n\n**Emil Ares** - Founded first company at 16, making six figures. Ex-Cambridge Physicist working on world's most powerful magnets (56MW) at the 1.4GW MagLab facility. Ex-UK Space Agency. Published researcher on AI for complex systems engineering design.\n\n**Gad Marconi** - Built and deployed on-site data center infrastructure at the 25MW Swiss Flumenthal hydroelectric plant. Ex-AI software engineer at Nuclear and Energy EPC. Ex-TU Delft.\n\n**Ask:**\n\nIf any of these describe you or your organization, we'd love to talk:\n\n* A neocloud, hyperscaler, developer or similar\n* A professional engineer working in data centers - architects, civil, mechanical, electrical; we’re hiring!\n\n---\n\nVisit to learn more! Contact us at [contact@marengox.com](mailto:contact@marengox.com) ", "url": "https://www.ycombinator.com/launches/SiK-marengo-accelerated-data-center-design", "votes": 41, "created_at": "2026-08-12T21:30:00.399Z" } ] }, { "id": 32985, "name": "Lumeria", "slug": "lumeria", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/5911b0640b4a7b63825ae466fa73c1fc1b24c0a3.png", "website": "https://lumeria.skin", "all_locations": "San Francisco, CA, USA", "long_description": "The Lumoscope is a multispectral camera that clips onto your phone and scans your skin at home with the same light used in clinics. You see what a mirror can't: a breakout days before it surfaces, and wrinkles changing over time. Our software then reads that history, connects flare-ups to causes you'd never suspect, and matches products to what your own data shows works. \r\n", "one_liner": "Oura Ring for Skin", "team_size": 3, "industry": "Healthcare", "subindustry": "Healthcare -> Consumer Health and Wellness", "launched_at": 1785008787, "tags": [ "Consumer Health Services", "Health Tech", "Health & Wellness", "Skincare", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare", "Consumer Health and Wellness" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": true, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/lumeria", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/lumeria.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://linkedin.com/lumeriaskin", "x_url": "https://x.com/LumeriaSkin", "fb_url": "https://www.facebook.com/profile.php?id=61591057680758", "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/5911b0640b4a7b63825ae466fa73c1fc1b24c0a3.png", "app_video_url": "https://bookface-images.s3.us-west-2.amazonaws.com/bf-public_profile/33043/app_video.mp4?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ASIAQC4NIECAK2XLVAEC%2F20261020%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20261020T022740Z&X-Amz-Expires=60&X-Amz-Security-Token=IQoJb3JpZ2luX2VjEHEaCXVzLXdlc3QtMiJFMEMCIAKxVeCuOoDsiv9OvW1mQ7mcS8gDT%2FRJKRw6KedW8MBiAh8pY6kF75O%2FWklLJYZpXjNpdv7D7g7DTGoNa63eX1RAKuUDCDoQABoMMDA2MjAxODExMDcyIgwvKBm%2Bb%2F8OodUhrswqwgODlPQnlHNviMegR2tgHBKcqrGQw%2BHtymSQ4Yp%2BHXvxgxZQmZdAto2k32TpvRz%2BVJE%2BqKCUOyrN3nF%2BYp8JTYAbk6GLOMZLQ8NNfzzBNWL8hlZzwhBWCB%2FK92gRfq%2BbLzMD4EZK5luTpQ73dNMYF5KWvVfVFb3JfvZdp6jebipH%2BOsZEvEVbPpI7RqhKxjIWCKuiisx5aGObeO7hsqRRnRVxOTbem6Emw3xwUMNtBgSeDAhJ584VF4e0rTl8AVT%2BPnpDr0OKvLbHZC2Ds%2F8AyDY6hIImlQE2UeauK%2B9ELmg6Yxp%2Fnz1kBNJ%2Fg2X9Jqi1JSLfvWbGkmuDAw89PFwEI0EfUUpXDyZbrY0b%2FSIPioJWPd5idMrJmtR9tYutaPns85TmjvYnh6TyjEksW1uIPi%2F%2FzTa2bGXAUVva1ybRBIV4TRnHjULbBAFlhE68zR2rLoypl%2BC0t0c1rrhWDfEV2%2BlqQRP38IJdPC5%2B32Q2eRMpzaalRIUh8mdZbHKmqaE2qczdAzfevEBI%2FtKPkLVhrYw2ND7bqysVVZRZYOE8rbkPsBRmpgp%2B7tz96eUMIbn3AaPWVW%2F1Au9bEDhTwoP1Kn6aTIw6P3a1gY6pwFIPtNtLcJWP8c793hYWjYfiWxa0eUSYLLcHApCHvGendn7NBYh6NUtrQ%2Fr4LAeOVtQUfme5Ilt3iCxJS%2FRGvvUbEz8jMx%2BOPIDNsEZvz5aI6VIlIz7els8agfAsUzZABhur3GtH9uqeXycsGWan43V8u0Ent37mQvBmvytxMaJCivuTGUjHlJquYXJoLnf%2BskSM1sq9omX%2Fxw7gT6pZpoO2xRdhnID%2FA%3D%3D&X-Amz-SignedHeaders=host&X-Amz-Signature=08fa5e5f5f79daa4aa81e425c02757a4425d1a11221aba8c26991eb8008d510c", "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 3452622, "full_name": "Maryanne Alhallak", "title": "Founder/CEO", "founder_bio": "Founder & CEO at Lumeria. Prev. Founder of 7-time award winning biotech startup (exited in 1.5 years). Beauty & lifestyle influencer (26k+ followers, 10M+ views in 2 months). Published author at 9 years old with <3 years of English knowledge as Syrian Refugee. UC Berkeley BS Bioengineering + BA Economics admit at 15. Prev. strategy work @ Flex, Pfizer, & PBS. Wetlab + CompBio Researcher.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/maryanne-alhallak-33b250241", "x_url": "https://x.com/malhalla06", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/83313ba7c25e91eda4e8d5c5f8b36be5038fab4b.jpg" }, { "user_id": 3470868, "full_name": "Anthea Guo", "title": "Founder/CTO", "founder_bio": "Founder and CTO at Lumeria. Previously Team Lead at Google, leading a 6-person Berkeley collaboration on DeepVariant. Built HPC pipelines at Boehringer Ingelheim. Built Streamind's phone-based pulse capture pipeline. Co-founded Computational Biology @ Berkeley.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/anthea-guo", "x_url": "https://x.com/theantheaguo", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/afe4281df1b70d51d9f9230998c91ac9550130e9.jpg" } ], "jobs": [], "news": [ { "title": "From Berkeley to Y Combinator: Lumeria Founders Build the Oura Ring for Skin - UC Berkeley Sutardja Center", "url": "https://scet.berkeley.edu/from-berkeley-to-y-combinator-lumeria-founders-build-the-oura-ring-for-skin/", "date": "Jul 01, 2026" } ], "launches": [ { "id": 109979, "title": "Lumeria - What if there was an Oura Ring for Skin?", "tagline": "What if there was an Oura Ring for Skin?", "body": "# Lumeria is the Oura Ring for Skin:\n\nTLDR: Lumeria is the Oura Ring for Skin: the Lumoscope clips onto your phone, scans your skin with the same multispectral light used in clinics, and Lumi, your AI skin coach, reads that history to personalize recommendations for you, so your routine is built on your data instead of someone else's.\n\nPreorders are open at the $199 founding price until **August 16. After that, the price only goes up**.  \n\n\n\n## ❌ The problem\n\nAnthea’s sister spent eight years chasing her eczema.\n\nThe cause turned out to be dust in the carpet.\n\nNo one caught it because no one was recording anything. Her skin was reacting to her sleep, her home, her products, the air around her, and none of it existed as data. Every dermatologist that took months to see started from zero. Every product she tried was a guess.\n\nThat is the **$190B** skincare industry in one story. It runs on guessing.\n\nLumeria ends that: it captures the missing data layer that gives you the power to understand and own your skin's health data, from home.\n\n![uploaded image](/media/?type=post&id=109979&key=user_uploads/3470868/08058377-6a96-414b-b03e-44ff36f8622a)\n\n## 🔬 How it works\n\nThe Lumoscope shrinks the multi-spectrum imaging dermatology clinics use into a clip for your phone. Press it against your skin and see yourself in three kinds of light: true color (RGB), UV, and cross-polarized.\n\n![uploaded image](/media/?type=post&id=109979&key=user_uploads/3470868/76776eef-52f9-4416-827c-36c74760f8dd)\n\nFor the first time, you see what your mirror can't: a breakout building days before it surfaces, sun damage hiding under flawless-looking skin, where inflammation is building up.\n\nEach scan becomes an objective read (hydration, redness, evenness, pigmentation, texture, pores) plus a score out of 100 and a skin age, so you can understand your progress.\n\nTrack them over time, and Lumi tells you exactly what to do. \n\n* It finds the cause, linking your breakouts to habits you'd never think of, from your sleep to your cycle to the pillowcase you forgot to wash. \n* It recommends products from a database of 18000+ products + ingredients, matched to what your scans actually show works.\n* It hands your dermatologist 90 days of images, trends, and context in one tap.\n* And it compounds: every scan adds to a history and increased understanding of your skin’s relevant context.\n\n![uploaded image](/media/?type=post&id=109979&key=user_uploads/3470868/c36570ca-4042-47bc-beb4-5f1a27014d6a)\n\n It works across every skin tone. It is private and encrypted by default. \n\n## 👥 Why we built this\n\nAnthea, our CTO, lived with this problem: suffering from eczema that would not heal, five months to reach a dermatologist, four misdiagnoses. She is an engineer who worked on computational biology at Google and Boehringer Ingelheim. Maryanne, our CEO, is a skincare creator with 25M+ views across her UGC portfolio, exited her first biotech company by 19, and still could not explain her own breakouts until Lumeria. We understand the helplessness people feel with their skin, so we built what was needed.\n\nWe are patent pending, shipping in Fall 2026, and built on 150+ interviews and 2 clinical partnerships with dermatologists and patients.\n\n![uploaded image](/media/?type=post&id=109979&key=user_uploads/3470868/d7da4795-da8c-4edd-8c96-58e540656067)\n\n## 🚀 Why preorder right now\n\nThis is the best moment there will ever be to own a Lumoscope:\n\n* **$199 until August 16.** The founding price is the lowest it will ever be, and it will go up to 329 otherwise. \n* **Founding preorders are first in line.** Units ship Sep - Oct 2026, ahead of general release.\n* **3 months of Premium free** plus a 30-day money-back guarantee.\n* **Founding member, not a ticket number:** a direct line to us and a real say in what we build next.\n* **Your skin data compounds from day one:** the earlier you set your baseline, the sooner your recommendations are built on your skin instead of averages.\n\nPreorder now at \n\nSee what Lumi has to say about your skin, and own your skin’s story today. ☀️\n\n💛 Reach out to us: [maryanne@lumeria.skin](mailto:maryanne@lumeria.skin) & [anthea@lumeria.skin](mailto:anthea@lumeria.skin)", "url": "https://www.ycombinator.com/launches/Sbr-lumeria-what-if-there-was-an-oura-ring-for-skin", "votes": 26, "created_at": "2026-08-12T14:00:00.255Z" } ] }, { "id": 32988, "name": "Markov", "slug": "markov", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/41a4a41e5482735389d2089a01ff2b862f56e307.png", "website": "https://www.markovstudios.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Markov builds expert computer-use datasets for frontier AI labs. Our datasets include screen recordings + mouse/keyboard actions of professionals using CAD, design and enterprise software.\r\n\r\nWe work with the top AI labs and we’ve gotten more than 350k downloads on HuggingFace. Our long-term goal is to be the full stack platform for computer-use AI, starting with data and then the infrastructure to train and deploy computer-use models.\r\n\r\nDev and Harish are technical co-founders. Dev studied aerospace engineering at IIT Madras and was the youngest team member at Sarvam - India’s top AI lab. Harish is an Emergent Ventures grantee and his prior robotics work was featured in top national publications.", "one_liner": "Expert computer-use data for frontier AI labs", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1782277833, "tags": [ "Reinforcement Learning", "Data Labeling", "Data Engineering" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/markov", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/markov.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/markov-artificial-intelligence/", "x_url": "https://x.com/markov_labs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/41a4a41e5482735389d2089a01ff2b862f56e307.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 1783684, "full_name": "Dev Mandal", "title": "Founder/CEO", "founder_bio": "Co-Founder & CEO at Markov. Prev at Sarvam, Aerospace engineering at IIT Madras", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dev-mandal-iitm/", "x_url": "https://x.com/DevvMandal", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5a53c70cbcddfae2c6c21e32faac5865be5702b2.jpg" }, { "user_id": 2921049, "full_name": "Harish Ashok", "title": "Founder/CTO", "founder_bio": "Co-Founder at Markov. Previously founded Zenith, a robotics company building AI-assisted hardware workflow tools. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/harish-ashok", "x_url": "https://x.com/harishashok", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fbfaea52b2a5bc846c87d04392062585acfb69ff.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109874, "title": "Markov - Data for computer-use AI", "tagline": "150K+ downloads across our HuggingFace datasets", "body": "# \\*\\*\\*\\*\n\n**TLDR;**\n\n* **Markov builds the data infrastructure for computer-use AI**, collecting synchronized screen recordings, mouse/keyboard actions, and annotations from real workflows.\n* **We specialize in hard-to-source expert data** **across CAD, architecture, engineering**, and other professional applications.\n* **We’ve already sold 15,000+ hours of data to frontier AI labs** and open-sourced leading computer-use datasets with **150k+ downloads on HuggingFace**.\n\n\n\n**The Problem:**.\n\nWhile frontier models have gotten much better at reasoning, they still struggle to reliably complete long, multi-step workflows across real desktop software. **The biggest bottleneck is the lack of high-quality training data.**\n\nTraining these models requires **synchronized screen recordings, mouse and keyboard actions, and high-quality annotations from people doing real work**. This is especially scarce for expert workflows in domains like CAD, architecture, engineering, finance, and other professional software.\n\n**That’s where we come in:**\n\nWe work with a vetted network of expert contributors to collect high-quality computer-use data across:\n\n* Consumer applications\n* Productivity software\n* Accounting and finance\n* CAD, architecture, engineering, and other expert workflows\n\n![uploaded image](/media/?type=post&id=109874&key=user_uploads/1783684/ba673ff4-e489-4753-b11b-4797762834fb)\n\n**So far we've:**\n\n* Sold more than 15,000 hours of data to frontier AI labs\n* Open-sourced the [**largest dataset of computer-use screen recordings**](https://x.com/DevvMandal/status/2032104265121140924?s=20)(**150k+ downloads on HuggingFace**)\n* Released the [**largest open gaming computer-use dataset**](https://x.com/DevvMandal/status/2072357115847671990?s=20)\n\n**Our Ask:**\n\nIf you're building computer-use agents and need high-quality training data - or have interesting workflows you'd like collected - we'd love to chat **:)**\n\nBook a meeting with us at Or shoot us an email at [founders@markovstudios.com](mailto:founders@markovstudios.com)", "url": "https://www.ycombinator.com/launches/SaA-markov-data-for-computer-use-ai", "votes": 11, "created_at": "2026-08-11T18:00:00.277Z" } ] }, { "id": 33012, "name": "Cosmic Robotics", "slug": "cosmic-robotics", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/610b3f44c95c65ea346121a48c57b8c81a3788ce.png", "website": "https://www.cosmicrobotics.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Our mission is to build a city on Mars. \r\n\r\nCosmic-1, the first equipment in our fleet, is an autonomous robot capable of heavy-lift manipulation. Its been deployed on some of the largest solar farms in the US, more than doubling labor productivity across tens of thousands of panel installs. \r\n\r\nWe're scaling in solar, expanding in data centers, and building lunar construction robotics under contract with NASA.", "one_liner": "Autonomous construction on Earth and beyond", "team_size": 8, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1783106253, "tags": [ "Hard Tech", "Robotics", "Construction", "Aerospace" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/cosmic-robotics", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/cosmic-robotics.json", "year_founded": 2024, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/cosmic-robotics/", "x_url": "https://x.com/cosmic_robotics", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/cosmic-robotics", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/610b3f44c95c65ea346121a48c57b8c81a3788ce.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 522502, "full_name": "James Emerick", "title": "Founder/CEO", "founder_bio": "Co-founder, CEO at Cosmic Robotics (YC S26). Previously robotics engineer at Built Robotics deploying autonomous earthmoving robots in production, and research engineer at Autodesk Research building adaptive platforms for construction. MS/BS in Mechanical Engineering from UPenn.", "is_active": true, "linkedin_url": "https://linkedin.com/in/jamesemerick", "x_url": "https://x.com/JamesAEmerick", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c3984e6d72a7fc6fad1faf754e9630152d5177fa.jpg" }, { "user_id": 403781, "full_name": "Lewis Jones", "title": "Founder/CTO", "founder_bio": "Co-founder, CTO at Cosmic Robotics (YC S26). Previously propulsion engineer at Relativity Space and prior experience at SpaceX, NASA JPL, and NASA JSC. Forbes 30 under 30. BS in Mechanical Engineering and Economics from Caltech.", "is_active": true, "linkedin_url": "https://linkedin.com/in/lewis-c-jones", "x_url": "https://x.com/thelcjones", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e307c9f2c5673760099031abcf82a9d8ea9dbff9.jpg" } ], "jobs": [], "news": [ { "title": "Cosmic Robotics' robots could speed up solar panel deployments | TechCrunch", "url": "https://techcrunch.com/2025/04/16/cosmic-robotics-is-building-robots-to-speed-solar-power-deployments-for-data-centers/", "date": "Apr 16, 2025" } ], "launches": [ { "id": 106899, "title": "Cosmic Robotics: Autonomous construction on Earth and beyond", "tagline": "Robotic construction equipment to build a city on Mars. We’re proving the technology by building large-scale infrastructure on Earth — utility-scale solar farms and hyperscale data centers.", "body": "Hey everyone 👋\n\nWe’re [James](https://www.linkedin.com/in/jamesemerick/) and [Lewis](https://www.linkedin.com/in/lewis-c-jones/) - co-founders of [Cosmic Robotics](https://www.cosmicrobotics.com). \n\n**TL;DR**\n\n**We build robotic construction equipment to build a city on Mars**. We’re proving the technology today by building large-scale infrastructure on Earth – utility-scale solar farms and hyperscale data centers. Our first machine, Cosmic-1, has already installed tens of thousands of solar panels in the field and more than doubled the productivity of traditional crews.\n\n \n\n**❌ The problem**\n\nAfter working at NASA JPL, SpaceX, and Relativity Space, we believe rockets are no longer the limiting factor to space exploration and that the first humans will land on Mars by 2034. **The new bottleneck is construction**. We won't be sending thousands of construction workers to Mars to build the energy, compute, and housing needed to sustain even a small city. **It has to be built by robots that don't exist yet**.\n\nOn Earth we have a similar problem - **we’ve lost the ability to build**. Construction productivity has declined for 50 years while manufacturing and agriculture became 10-15x more productive through automation. **The missing technology is robotic manipulation for the tasks still done by hand**: lifting, placing, fastening.\n\nMeanwhile, there is a generational infrastructure buildout happening right now. **$7 trillion** is going into AI data centers, power, and compute over the next six years, **more than the Apollo Program, Manhattan Project, and Interstate Highway System combined**. 40% of US data center projects are running behind schedule, and the skilled labor shortage is the #1 cause. \n\n**🤖 What we've built**\n\nCosmic-1 is a 10,000 lb autonomous robot capable of heavy-lift manipulation. In solar, it picks up 90 lb panels and places them onto racking with millimeter precision in any environment: 100-degree heat, freezing rain, snow, and dust. It seamlessly integrates into existing construction processes. Across tens of thousands of cycles it has more than doubled the labor productivity of solar installation, on some of the largest solar projects in the US.\n\n![uploaded image](/media/?type=post&id=106899&key=user_uploads/522502/bb5f86ec-a23b-4ffd-a173-87f36ce6fb1c)\n\n**👬 The Team**\n\nWe come at this problem from two angles. James (CEO) grew up on construction sites, then spent his career bringing robots to the field with Built Robotics and Autodesk Research. He learned you don't change the oldest industry in the world by disrupting it; you earn your way in by giving crews better tools.\n\nLewis' (CTO) life ambition is to build a city on Mars. His work at NASA, SpaceX, and Relativity Space spans rocket engines, launch vehicles, satellites, space stations, and, yes, Martian robots. When it became clear the space industry was well on its way to Mars, he went looking for the construction robots. When he didn't find them, we started Cosmic.\n\nWe’ve brought together a team of industry experts with experience building robots, rockets, drones, self-driving cars, humanoids, and nuclear reactor operations. We know what it takes to get a machine to show up day-in-day-out on the jobsite.\n\n![uploaded image](/media/?type=post&id=106899&key=user_uploads/522502/87f30a8e-6f83-490e-8bef-04225d1a525e)\n\n**🚀 What's next**\n\nWe're scaling in solar, taking Cosmic-1 into data center construction, and working with NASA on lunar construction robotics. Revenue and data from deployments on Earth train the systems we'll deploy on Mars.\n\n**🙏 Our asks**\n\n1. **Know a hyperscaler or neocloud construction team?** Please connect us.\n2. **Share this post** and [follow our journey](https://www.linkedin.com/company/cosmic-robotics/).\n\n👉 Contact us at [james@cosmicrobotics.com](mailto:james@cosmicrobotics.com) | [cosmicrobotics.com](http://cosmicrobotics.com)", "url": "https://www.ycombinator.com/launches/RoB-cosmic-robotics-autonomous-construction-on-earth-and-beyond", "votes": 28, "created_at": "2026-07-23T17:36:46.436Z" } ] }, { "id": 33013, "name": "Operon", "slug": "operon", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/aaeb7f2ff7aa5b0b7e64888230be2a7e46342f66.png", "website": "https://operonsolutions.com/en", "all_locations": "San Francisco, CA, USA", "long_description": "Operon is the data intelligence layer for heavy industry. We ingest the messy documents a plant runs on (P&IDs, schematics, datasheets, scans) and turn them into one structured, queryable model of the facility.\r\n\r\nOn top of it, engineers query their plant in natural language and build agentic workflows for safety, compliance, and design. \r\n\r\n", "one_liner": "Agentic data layer for manufacturing & process industries. ", "team_size": 0, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1782593141, "tags": [ "Computer Vision", "Manufacturing", "Infrastructure" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/operon", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/operon.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/operonsolutions/?viewAsMember=true", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/aaeb7f2ff7aa5b0b7e64888230be2a7e46342f66.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 2155140, "full_name": "Anderson Chen", "title": "Founder/CEO", "founder_bio": "Founde/CEO @ Operon Solutions.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/anderson-chen-2b6941216/", "x_url": null, "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/289ff08d5218c7b14ea8d2d173dfb9330c17e927.jpg" } ], "jobs": [ { "id": 96232, "title": "Founding FDE", "url": "https://www.ycombinator.com/companies/operon/jobs/w7qf3HJ-founding-fde", "location": "San Francisco, CA, US / Remote (US)", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$100K - $150K", "equity_range": "0.20% - 1.00%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [] }, { "id": 33014, "name": "Inkbox", "slug": "inkbox", "former_names": [ "Inkbox (Vectorly Inc)" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/59d6214df81a05c739db52287017ae33c8c45f0d.png", "website": "https://inkbox.ai", "all_locations": "San Francisco, CA, USA", "long_description": "For your AI agents to behave like teammates, we give them persistent identities with their own email, phone, and iMessage. We also provide them an agent-to-agent channel, so they can communicate for task delegation.", "one_liner": "The identity and communication layer for AI agents", "team_size": 3, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1782881612, "tags": [ "Developer Tools", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/inkbox", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/inkbox.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/inkbox-ai", "x_url": "https://x.com/inkbox_ai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/59d6214df81a05c739db52287017ae33c8c45f0d.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 778892, "full_name": "Ray Liao", "title": "Founder/CEO", "founder_bio": "PhD in Computer Science from MIT. Previously cofounder/CTO at Empallo, an AI-powered virtual clinic for cardiovascular care.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ruizhi-liao/", "x_url": "https://x.com/rayruizhiliao", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/af788a5cd4056a418504ce1f6575ee97d05e4f22.jpg" }, { "user_id": 2770188, "full_name": "Dima Vremenko", "title": "Founder", "founder_bio": "Co-founder at Inkbox. Previously Research Associate at Harvard Medical School, where he built ML pipelines for clinical prediction models using gigapixel pathology images and EHRs, co-authored 6 publications, and contributed to an R01 NIH Grant. BS in Biochemistry and Computer Science from Northeastern.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dimavrem/", "x_url": "https://x.com/DVremenko", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/03bf2d678fb1434af863ffe50748028f1b1ddf01.jpg" }, { "user_id": 2773563, "full_name": "Alex Wilcox", "title": "Founder/CTO", "founder_bio": "Cofounder at Inkbox. Previously ML engineering and computer networking research. MS in CS and dual BS in Math and Econ from Northeastern University.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/alxwlcx", "x_url": "https://x.com/alxwlcx", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/aed3f2339877d7f84f73c49b3b73598cee41245b.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 104960, "title": "Inkbox: Give your AI agents email, phone, iMessage, and an internet address", "tagline": "The identity and communication layer for AI agents", "body": "**TL;DR:** Inkbox is the identity and communication layer for AI agents: giving agents their own email, phone, iMessage, and internet address. Inkbox pulls together what your agent sees across email, calls, iMessage, and more, so your agent can act on the full context.\n\nHey YC! We're Ray, Alex, and Dima building Inkbox.\n\n**How we got here**\n\nA few months ago we were building web-automation agents. Every time an agent needed to log into a portal, it had to receive a confirmation email or get a 2FA code. And when we were away from our desks, we wanted to seamlessly contact the agents themselves: text them, check in, get a reply. We ended up stitching together services like Gmail, Twilio, and ngrok: separate services, each built for humans, none of them sharing any identity or context. \n\nThat's the problem we set out to fix. We didn't need another email or voice API. We needed one service to tie its identity together with rich context. So we built it.\n\n[**Launch video**](https://www.youtube.com/watch?v=oJKl6tPHxRA)\n\n**What Inkbox does**\n\nOur API gives every agent:\n\n* A unique email inbox (+ custom domain)\n* A phone number that can call and text\n* iMessage\n* A persistent internet address (HTTPS tunnel) to receive internet traffic, like webhooks\n* An encrypted credential vault with TOTP/2FA\n\n**Who we are**\n\n* Ray (CEO): CS PhD from MIT; previously cofounded Empallo, an AI virtual clinic for cardiovascular patients.\n* Alex (CTO): MS in CS @ Northeastern. Previously ML engineering and computer networking research.\n* Dima (CAIO): built ML pipelines on gigapixel pathology images at Harvard Medical School; co-author on 6+ peer-reviewed publications.\n\n[**Quick demo with Claude**](https://www.youtube.com/watch?v=MTVkW_EG-Y4): Watch Claude Code provision an email, phone, and iMessage in a Minute.\n\n**Our ask**\n\n* **Try it**. Drop Inkbox into your preferred agent harness: we have plugins for[ Hermes](https://github.com/inkbox-ai/hermes-agent-plugin),[ OpenClaw](https://github.com/inkbox-ai/openclaw-plugin),[ Claude Code](https://github.com/inkbox-ai/claude-code-plugin), and[ Codex](https://github.com/inkbox-ai/codex-plugin). Each one runs a wizard that provisions Inkbox resources for your agent.\n* **Ask your agent to self signup**. _Sign up for an Inkbox email account by following the guide at[ https://inkbox.ai/docs/get-started/agent-signup.md](https://inkbox.ai/docs/get-started/agent-signup.md) and refer to[ https://inkbox.ai/llms.txt](https://inkbox.ai/llms.txt) for full documentation._\n* **Intros**. We'd love intros to teams running fleets of agents that each need their own identity to act in the real world: reaching out to people and businesses (email/phone/iMessage), holding their own accounts, and authenticating with a built-in credential vault.\n\nContact Ray at [ray@inkbox.ai](mailto:ray@inkbox.ai). Would love your feedback.", "url": "https://www.ycombinator.com/launches/RIu-inkbox-give-your-ai-agents-email-phone-imessage-and-an-internet-address", "votes": 21, "created_at": "2026-07-08T00:25:54.653Z" } ] }, { "id": 33016, "name": "Gamgee", "slug": "gamgee", "former_names": [ "Gamgee Technologies" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/6f6bf83f60f82f47f1a3b961153608810edebb3b.png", "website": "https://gamgee.io/", "all_locations": "San Francisco, CA, USA; Remote", "long_description": "​Gamgee began with Rosie, founder Paul Conyngham’s dog. After surgery, chemotherapy and immunotherapy failed, Rosie was given months to live. Paul used AI and genomic analysis to identify targets in her tumour, then worked with scientists at UNSW to produce what is believed to be the first fully computationally designed mRNA cancer vaccine delivered to a dog. Following treatment, several tumours shrank significantly and her quality of life improved\r\n\r\nGamgee is building an end to end platform for personalised mRNA cancer vaccines for dogs. We sequence a dog’s tumour and healthy DNA, identify mutations unique to its cancer, and design a vaccine that teaches the immune system what to attack. We manage the entire process- from vet intake and tissue collection through genomic analysis, vaccine production, treatment and response monitoring\r\n\r\n​We are now turning that one off effort into a repeatable clinical system. Gamgee is running clinical trials in Australia, working with vets, oncologists and specialised laboratories, and accepting cases globally. Our goal is to make personalised cancer treatment accessible to every dog that needs it- and use each case to improve the treatment for the next", "one_liner": "Curing cancer. Starting with dogs.", "team_size": 2, "industry": "Healthcare", "subindustry": "Healthcare -> Drug Discovery and Delivery", "launched_at": 1784748212, "tags": [ "Biotech", "Genomics", "Oncology", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare", "Drug Discovery and Delivery" ], "regions": [ "United States of America", "America / Canada", "Remote", "Fully Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/gamgee", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/gamgee.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/109672135", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/6f6bf83f60f82f47f1a3b961153608810edebb3b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 564809, "full_name": "Paul Conyngham", "title": "Founder/CEO", "founder_bio": "Founder at Gamgee Technologies (YC S26), building personalised mRNA cancer vaccines for dogs.\n\nGamgee grew out of my work using AI and genomics to help my dog Rosie fight cancer. \n\nPreviously co-founded Core Intelligence, building applied AI systems across startups and enterprise, and helped grow Australia’s AI community as a founding member of the Data Science & AI association of Australia. \n\nBackground in ML, data science and electrical and computer systems engineering", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/paustevenlconyngham/", "x_url": "https://x.com/paul_conyngham", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4fda91993052b2e444cfffa9d59e27aa42567028.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109527, "title": "Gamgee: Personalised mRNA cancer vaccines for dogs", "tagline": "Input the tumour. Output the vaccine. 300+ dogs on the waitlist.", "body": "**We believe that AI meets biology is the biggest opportunity of the 21st century.** We are using AI to build autonomous systems that read biological failure, discover how to control it, and design precise repairs. We are starting with the first computationally personalised mRNA cancer vaccine for dogs.\\\n\\\n**My dog Rosie was dying of cancer.** I designed her an mRNA vaccine. Several of her tumours shrank 50-75%. I am not a biologist - I spent 17 years in machine learning. Doing it by hand took a year and around $20,000. Gamgee is turning that process into software. Our target is a vaccine design in 48 hours and an end-to-end cost of $2,000\n\n\n\n**How it works**\n\n* We **send you a collection kit** to take to your vet or veterinary oncologist. Your vet collects the tumour and matched healthy samples, which are returned for DNA and RNA sequencing.\n* Our software finds and ranks the mutations worth targeting, and keeps the evidence behind every call.\n* **Out comes an mRNA vaccine** design ready for manufacture. Gamgee designs the vaccine and oversees and coordinates every step through to administration by the treating veterinarian.\n\n\\\n**The market**\n\nApproximately **six million dogs** are **diagnosed with cancer** every year in the United States, and cancer is the leading cause of death in adult dogs. Owners who pursue specialist treatment spend thousands on surgery, chemotherapy and radiation.\\\n\\\n\\\n**The compounding data advantage**\n\nEvery dog we treat returns sequencing, immune and outcome data for a species where datasets remain scarce. That data improves target selection, which improves the next vaccine, which produces better data. We are not just early to the problem - we are the ones accumulating the data that solves it.\n\n\\\n**Proof so far**\n\n* Full computational pipeline built and executed end-to-end as a non-biologist using AI\n* 50-75% shrinkage in the responsive tumours. \n* Microscopic analysis found substantial IBA1+ macrophage and CD3+ T-cell infiltration in Rosie’s responsive lesions, with little in the non-responsive tumour.\n* Second-generation vaccine completed and ready\n* Over a billion impressions on Rosie's story.\n* Backed by Y Combinator (S26) \n\nRosie was one patient. Our next step, already well underway, is a clinician-led trial in 30 dogs to evaluate safety, immune response and clinical outcomes.\n\n\\\n**Why dogs first, and not only dogs**\n\nMost modern dog breeds were created within the last few hundred years from small founder populations and then bred within closed gene pools. This concentrates inherited cancer risk, leaving many breeds unusually prone to particular cancers.\\\n\\\nPet dogs develop cancer naturally, not in a laboratory. Their immune systems are intact and their disease progresses on timelines that let us learn quickly. Dogs are a real market in their own right and the fastest place to get good at personalised cancer vaccines.\n\n\\\n**Our ask**\n\n* **If your dog has cancer, please do get in touch about our upcoming clinical trial.**\n* Veterinary oncologists and specialty hospitals who want to be a trial site.\n* Investors in longevity or personalised medicine - would rather meet you before the trial reads out than after.\n\nWe are building Gamgee so that no other dog waits for hope which arrives too late.\n\nReach me at [paul@gamgee.io](mailto:paul@gamgee.io) or [gamgee.io](http://gamgee.io)\\\n\\\nI love you Rosie.", "url": "https://www.ycombinator.com/launches/SUZ-gamgee-personalised-mrna-cancer-vaccines-for-dogs", "votes": 186, "created_at": "2026-08-11T17:00:00.187Z" } ] }, { "id": 33017, "name": "ProvenMetal", "slug": "provenmetal", "former_names": [ "Syncra" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/450cf955a0b5a1e301afe718484dd30ffb22fb67.png", "website": "https://provenmetal.com", "all_locations": "San Francisco, CA, USA", "long_description": "ProvenMetal manufactures the fastest circuit boards in the United States.\r\n\r\nDomestic circuit board manufacturing is slow and expensive, and while reshoring is happening, the supply chain isn't there yet.\r\n\r\nWe automate the bottlenecks - part procurement, DFM review, and assembly - delivering circuit boards in days instead of weeks.\r\n", "one_liner": "AI-native electronics manufacturing for the American industrial base", "team_size": 4, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1783632830, "tags": [ "Hardware", "Manufacturing", "Electronics" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/provenmetal", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/provenmetal.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://linkedin.com/company/provenmetal", "x_url": "https://x.com/provenmetal", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/provenmetal", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/450cf955a0b5a1e301afe718484dd30ffb22fb67.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 3677443, "full_name": "Johnny Doyle", "title": "Founder/CEO", "founder_bio": "Co-Founder & CEO at ProvenMetal (YC S26).\nPreviously shipped IoT building management systems. \nDropped out of Engineering with Management @ Trinity College", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/johnny-doyle-69022718a/", "x_url": "https://x.com/jdoyleirl", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e0f543b094d851454fd151bf153f573591886fef.jpg" }, { "user_id": 1001813, "full_name": "Will Carkner", "title": "Founder/CTO", "founder_bio": "Co-Founder at ProvenMetal (YC S26). Previously Electrical Engineer at Lumindt working on hydrogen battery systems, studied Electrical Engineering at Trinity College Dublin, and researched metagenomics at ICL.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/will-carkner-518b77228/", "x_url": "https://x.com/CarknerWill", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9e2e5b2da14ac447cd242e0bc67bf50701494890.jpg" } ], "jobs": [ { "id": 109684, "title": "Founding Engineer", "url": "https://www.ycombinator.com/companies/provenmetal/jobs/rBRC3Vu-founding-engineer", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$100K - $200K", "equity_range": "1.00% - 4.00%", "experience": "Any (new grads ok)", "visa": "Will sponsor", "skills": [] } ], "news": [], "launches": [ { "id": 105450, "title": "ProvenMetal - Fast-turn, American made PCBs", "tagline": "Ship hardware just as fast as software", "body": "We're Johnny and Will - founders of ProvenMetal (YC S26). We assemble the fastest circuit boards in the US, out of a garage in Dogpatch, San Francisco.\n\nSend us your design files and get assembled boards back in days instead of weeks.\n\n\n\n**The Problem**\n\nHardware companies - start-ups and large manufacturers alike - need speed to survive, and right now they wait weeks for circuit boards.\n\nThat time is lost to quoting, design review, part procurement, and assembly - bottlenecks we're automating.\n\nAt the same time, the reshoring wave is driving demand to manufacture critical infrastructure - drones, robotics, defense, energy - domestically.\n\nBut the capacity to assemble boards domestically, quickly and at scale, isn't there. Decades of offshoring hollowed it out, and what's left is slow and expensive.\n\nThe supply chain needed to revitalise American electronics manufacturing hasn't been built yet. We're building it.\n\n**What We Do**\n\nWe're removing the bottlenecks that stand in the way of fast domestic PCB manufacturing.\n\nSend us your design files, get instant quoting and rapid design review.\n\nWe source every component, assemble in San Francisco, and X-ray every board - delivered in days instead of weeks.\n\n**The Founders**\n\nWill Carkner (CTO) - Electrical Engineer, previously shipped the PCB testing pipeline at Lumindt.\n\nJohnny Doyle (CEO) - Mechanical and Manufacturing Engineer, previously built IoT building management systems.\n\n**The Ask**\n\nIntros to drone companies that need hardware fast.\n\n![uploaded image](/media/?type=post&id=105450&key=user_uploads/3677443/8207a1f2-8f6c-41a2-ade8-66fa8d5f7c59)\n\n![uploaded image](/media/?type=post&id=105450&key=user_uploads/3677443/75079b59-06bb-49f5-898f-dff57e93579a)\n\n![uploaded image](/media/?type=post&id=105450&key=user_uploads/3677443/eeabd5ee-fa61-411c-97da-519125e27c0d)\n\n", "url": "https://www.ycombinator.com/launches/RQo-provenmetal-fast-turn-american-made-pcbs", "votes": 34, "created_at": "2026-07-13T16:18:44.326Z" } ] }, { "id": 33018, "name": "Agentcard", "slug": "agentcard", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/5470e0782b6543d28db5fba3d76fabbdb0b65435.png", "website": "https://www.agentcard.com", "all_locations": "San Francisco, CA, USA", "long_description": "Agentcard let's companies buy anything online. We work with companies building personal assistants to enable payments.", "one_liner": "let agents buy anything online", "team_size": 4, "industry": "Fintech", "subindustry": "Fintech", "launched_at": 1780682746, "tags": [ "Fintech" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/agentcard", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/agentcard.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "https://x.com/agentcardhq", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/5470e0782b6543d28db5fba3d76fabbdb0b65435.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 347376, "full_name": "Karen Serfaty", "title": "Founder/CEO", "founder_bio": "CEO & founder at Agentcard. Previously CEO at Atlas, an expense card for global teams (acq. by Remote.com). GM, Cards at Remote.com. Co-founder/CEO at Palabra. Co-founder/CTO at Nilus. Studied Universidad de San Andrés.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/keyserfaty/", "x_url": "https://twitter.com/keyserfaty", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9a7f149b222e50cb6a5d45e0924b01e869722853.jpg" }, { "user_id": 363364, "full_name": "Felipe Abello", "title": "Founder/CTO", "founder_bio": "cofounder at agentcard.sh - prev: cofounder at infactory.ai - founder & CEO at Nara (Helloguru W22) - ex @Rappi (1st employee).", "is_active": true, "linkedin_url": "https://linkedin.com/in/pipe-abello-2512b9195", "x_url": "https://twitter.com/PipeAbellos", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/93a0a14dff5fd9699028e53e41ccea1045766a8c.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112244, "title": "Agentcard: Cards for AI agents", "tagline": "Let agents buy anything online", "body": "[ ](https://www.youtube.com/watch?v=RaeATYNiBbY)\n**TL;DR**\\\nBuying online with agents is hard. You can't share your card with your agent, browser automations are broken and slow, and most checkout pages block agents, mistaking them for bots. Agentcard gives companies a solution for storing users' credit cards and completing checkout pages with a 99% success rate.\n\n**Ask**: building an agent that buys things on behalf of your users? send us a msg: [founders@agentcard.sh](mailto:founders@agentcard.sh)\n\n![uploaded image](/media/?type=post&id=112244&key=user_uploads/347376/2d085793-25d2-412b-9d20-40785159d157)\n\n**Problem**\n\nCompanies building personal AI assistants want to give their users a simple way to buy using agents. Today, they need to 1) create browser automations to navigate online stores, 2) store users' cards, and 3) complete checkout pages. None of this is their main product. And current solutions don't work: agentic protocols for storing cards have an 8% card acceptance rate (i.e., most of your cards don't work) and are US-only; agentic browsers get blocked when trying to complete a checkout page; companies store cards without being PCI compliant; and, finally, any dispute becomes a problem for the merchant.\n\n**Our solution**\n\n* **Card vault**: a PCI-compliant way to store users' cards.\n* **Agent-browser integration**: native integration with the main agent browsers.\n* **Bring your own browser**: simple integration with your own browser.\n* **Checkout completion**: 99% success rate on the most popular merchants (Shopify, Stripe and others).\n\n**How it works**\n\n![uploaded image](/media/?type=post&id=112244&key=user_uploads/347376/9b3b8719-6d1d-44c1-9995-685b7e2d0036)\n\n1. **Store card**: share a safe link with your customer and store their cards with Agentcard.\n\n![uploaded image](/media/?type=post&id=112244&key=user_uploads/347376/5282b31a-6d30-4a8e-a00d-972dfde54697)\n\n2. **Shop**: with the agent browser of your choice, complete the purchase on behalf of your user using a dummy card.\n\n![uploaded image](/media/?type=post&id=112244&key=user_uploads/347376/612a72b0-6f23-4791-a731-2acc837afe5b)\n\n\n3. **Human authorization & checkout handoff**: share the cart URL with Agentcard and let us complete it.\n\n![uploaded image](/media/?type=post&id=112244&key=user_uploads/347376/de5efb81-bc1a-4caa-ba18-fe25af0f257f)\n\n4. **Success**: we notify you when the purchase is done.\n\n**Team**\n\n![uploaded image](/media/?type=post&id=112244&key=user_uploads/347376/d538ba37-4d5d-4693-a2cb-269790b34248)\n\n* **Karen Serfaty**: CEO @ Agentcard. Prev. CEO @ Atlas (acq. by Remote). An expense card for global teams. Before that CTO @ Nilus incubated at Harvard; software engineer for +10 years.\n* **Felipe Abello**: CTO @ Agentcard. Employee #1 @ Rappi. Prev. CEO @ Helloguru (YC W22). Prev. co-founder @ Infactory.\n* **Braxter**: Sales @ Agentcard. All things Agentcard. Haters will say he’s an AI character. Don’t listen.", "url": "https://www.ycombinator.com/launches/TCO-agentcard-cards-for-ai-agents", "votes": 38, "created_at": "2026-08-26T18:00:00.618Z" }, { "id": 103608, "title": "Agentcard: Debit cards for AI agents", "tagline": "Everything an agent needs to pay for things online", "body": "hey everyone, we're [Karen](https://x.com/keyserfaty) and [Felipe](https://x.com/PipeAbellos), and we are building [Agentcard](https://www.agentcard.sh/): debit cards for AI agents.\n\n# TL;DR\n\nAI agents are finally smart enough to do real things for you. but the second they need to pay for something, they're stuck, and handing an agent your real card (no limits, no controls) is a terrible idea.\n\nso we built [Agentcard](https://www.agentcard.sh/): give your agent its own debit card: fund a wallet, set a budget, and issue one-time cards that auto-close after a single charge. works with any agent over MCP, the CLI, or our API. its everything an agent needs to pay for things online. here's the lineup:\n\nso we built Agentcard — everything an agent needs to pay for things online. here's the lineup\n\n# the team\n\n[Karen](https://x.com/keyserfaty) (CEO): former CEO @ Atlas (acquired by Remote), then turned down a job at Stripe to build Agentcard instead.\n\n[Felipe](https://x.com/PipeAbellos) (CTO): second-time YC founder (HelloGuru, W22) and first employee at Rappi (YC W16), where he shipped consumer products at scale. \n\n# the problem\n\nagents can now research, plan, and decide but the second they need to pay, they're stuck. the workarounds are all bad: hand your agent your real card and you've got no limits and no controls (one bad prompt from disaster), or build your own issuing + checkout stack and lose months becoming a payments company. so agents do the first 90% and hand checkout back to you, the most valuable step is the one they can't take.\n\n# the solution\n\n[Agentcard](https://agentcard.sh/) is everything an agent needs to pay for things online:\n\n* **💳 cards**\\\n give any agent its own debit card: fund a wallet, set a budget, and issue one-time cards that auto-close after a single charge. if something goes wrong, the blast radius is one capped card, not your bank account.\\\n \n* **🛒 buy (new — launching today!)**\\\n you don't even wire up a merchant. connect once, then tell your agent \"buy me sushi\" and it runs the whole thing — search, cart, checkout, payment — in the conversation. first up: DoorDash, right from inside Claude. it feels like magic 🥹\\\n \n* **🏢 Agentcard for companies**\\\n drop [Agentcard](https://www.agentcard.sh/) into your own product and give your users' agents their own cards at scale, with budgets and spend controls built in.\n\nall of it works with any agent: over MCP, the CLI, or our API. and people are already using it: agents out there paying for OpenAI subscriptions, groceries and infra on their own.\n\n# our ask\n\n* building with agents? try it → [agentcard.sh](http://agentcard.sh), and tell us which merchant we should add next\n* building a product whose agents should buy things? let's talk → [founders@agentcard.sh](mailto:founders@agentcard.sh)\n\nthanks everyone 🙏", "url": "https://www.ycombinator.com/launches/Qx6-agentcard-debit-cards-for-ai-agents", "votes": 31, "created_at": "2026-06-25T19:25:46.736Z" } ] }, { "id": 33019, "name": "Agnost AI", "slug": "agnost-ai", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/e1595f92ea75b1f367addedb5703852b1c7efded.png", "website": "https://agnost.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Agnost AI reads every conversation a company's AI Agent has with its users to find where it's silently failing and then turns that same data into custom models that run the agent better, faster & cheaper than frontier models.", "one_liner": "We find where AI agents fail, then train better models to run them.", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Analytics", "launched_at": 1783481375, "tags": [ "Developer Tools", "Machine Learning", "Analytics", "Monitoring", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Analytics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/agnost-ai", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/agnost-ai.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://linkedin.com/company/agnostai", "x_url": "https://x.com/agnostai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/e1595f92ea75b1f367addedb5703852b1c7efded.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/b09996d90ed38a4b5d62581651fffd7e0885e6b2.png" ], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 1987056, "full_name": "Shubham Palriwala", "title": "Founder/CEO", "founder_bio": "Co-founder and CEO of Agnost AI.\nYoungest Engineer on Cisco's Analytics team, First Hire at Formbricks (Open-Source Qualtrics), wrote code in Bitcoin & other open-source projects across OWASP and the Linux foundation.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shubhampalriwala/", "x_url": "https://x.com/ShubhamInTech", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/16f7d9652057acfedcb3ad0c94ecab3e15758aba.jpg" }, { "user_id": 2055517, "full_name": "Parth Ajmera", "title": "Founder/CTO", "founder_bio": "Co-Founder/CTO at Agnost AI. Graduated in Computer Science from IIT Madras, ranking 159 among over 1 million candidates. Previously led graphics engineering at Infurnia and built terabyte scaled data pipelines at Microsoft", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/parthajmera/", "x_url": "https://x.com/prrthh132", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/93dc08b7da172c3288f1f798f2b9a0b27db75f16.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111310, "title": "Agnost AI: Turn Your Agent Traces Into a Faster, Cheaper, & more Accurate Custom Model", "tagline": "We use your agent’s production traces to train a specialized model that performs better, responds faster, and costs less.", "body": "Your agent logs aren’t just for debugging. They’re your training data.\\\nHey YC! We’re Shubham and Parth, childhood friends and founders of Agnost AI.\n\n**TL;DR**: Agnost AI uses your agent’s production traces to train a model specialized for your exact workload. Our first custom model for our customer delivered them a:\n\n* 22.9% higher task success,\n* 90.2% lower median latency, and\n* 94.5% lower cost than Opus 4.8.\n\nDemo: \n\n### What we found\n\nA few weeks ago, we launched Agnost AI as product analytics for AI agents. We read conversations between users and agents to uncover failures, behavior drift, hallucinated links, frustration, and churn signals.\n\nWe expected to find bugs. Surprisingly, we also found a lot of infrastructure waste. Many customer-facing agents use frontier models for narrow, repetitive jobs:\n\n* Route a request to the correct tool\n* Have a narrow system prompt with the given inputs\n* Retrieve the relevant data from internal tool calls\n* Parse & format the final response\n\nThe underlying frontier model can write code, solve advanced mathematics, understand biology, and reason across hundreds of domains. **Your agent uses a small, repeatable slice of that capability while paying frontier-model prices for every request.**\n\n### We trained a specialist\n\nWe used historical production traces to train our first workload-specific model for an early customer: **agnost-<redacted>-0.1**\n\nWe evaluated it against Opus 4.8 on 780 held-out customer traces for an Identifier Extraction Agent. The results:\n\n* Task success: 71.5% → 87.9%, 22.9% relative improvement\n* Median latency: 4.80s → 0.47s, 90.2% lower\n* Tail latency: 11.20s → 1.35s, 87.9% lower\n* Cost per 1,000 completed tasks: $42.00 → $2.30, 94.5% lower\n\nThis is one model for one workload, not a claim that specialist models should replace frontier models everywhere.\n\nFrontier models remain the right choice for open-ended work. But once an agent’s production job becomes predictable, **paying for general intelligence on every request stops making sense**.\n\n### How it works\n\n1. Send your existing traces to Agnost through OpenTelemetry or our SDK.\n2. We identify the recurring workflows your agent performs.\n3. We create an eval set from held-out production traces.\n4. We train a specialist model on the historical workload.\n5. We benchmark it against your current model on success, latency, and cost.\n\nYou only consider switching when the specialist wins on your own workload. Your logs go from debugging data to evaluation data and finally to training data recursively.\n\n### Who we are\n\nWe’re childhood friends, eight years and counting.\n\n* Shubham: Youngest engineer on Cisco’s analytics team, first hire at Formbricks, and contributor to Bitcoin, OWASP, and the Linux Foundation.\n* Parth: IIT Madras CS, ranked 159 out of 1.5 million. Built Spark pipelines at Microsoft and led GPU technology at Infurnia.\n\n### Our ask\n\nIf you have meaningful daily traffic, send us an anonymized trace export. We’ll analyze the workload, build a held-out evaluation set, and tell you whether a specialist model is worth training. If it is, we’ll benchmark it against your current frontier model.\n\nEmail me at [shubham@agnost.ai](mailto:shubham@agnost.ai) or book a call at [call.agnost.ai](http://call.agnost.ai)\n\nMore about us at: [agnost.ai](https://agnost.ai)\n\n![uploaded image](/media/?type=post&id=111310&key=user_uploads/1987056/6599ed29-ddf9-4d02-9232-425af1535be7)\n\n", "url": "https://www.ycombinator.com/launches/SxK-agnost-ai-turn-your-agent-traces-into-a-faster-cheaper-more-accurate-custom-model", "votes": 97, "created_at": "2026-08-20T16:00:00.252Z" }, { "id": 107260, "title": "Agnost AI: Product Analytics for AI Agents", "tagline": "We read every conversation your users have with your agents & chatbots and surface hidden insights & product feedback. We are working with teams at Google, Exa and Corgi", "body": "**Your users are telling your agent exactly why they churn! You’re not listening.**\n\nHey YC! We’re Shubham and Parth, childhood friends, founders of **Agnost AI**.\n\n**TL;DR: Agnost AI reads every conversation your chat or voice agent has with your users. We show you what your users want, where they get frustrated, and why they leave.**\n\n**The feedback hiding in your logs becomes discoveries, evals, and fixes ready to ship. We’re already processing over 1 million events every day and growing.**\n\nTry us out at[ agnost.ai](http://agnost.ai)\n\n### \n\n### **The problem**\n\nTraditional product analytics were built for clicks.\\\nObservability was built for token costs, latencies, and deterministic errors.\n\nBut every product is rapidly moving from UI to conversation, and agent failures don’t look like traditional failures. The LLM returns 200 OK and says “sorry, I can’t do that.” The real failures are hiding in your conversations:\n\n* “can it remember this for next time?” → feature request\n* “you said it’s done, but nothing happened” → bug\n* “i’ve told you this three times already” → frustration\n* “nvm i’ll just do it myself” → churn\n\nToday teams throw more evals at this. But evals only tell you whether you fixed something, not what to fix. You can’t write a test for a problem you haven’t discovered yet.\n\n### **How it works**\n\n1. **Install our skill:** 3 lines of code, or point your OpenTelemetry exporter at us. Any LLM, any framework. Add any custom metadata like user tier, model versions, etc. We’re at:[ docs.agnost.ai](http://docs.agnost.ai)\n2. **Every conversation becomes a signal:** feature requests, bugs, frustration, churn risk, failed upgrades. Detected for your product, with the exact users & conversations behind each.\n3. **The insight becomes action:** when hundreds of users rageprompt at your agent due to a specific bug, Agnost AI hands you the insight with the exact conversations behind it: \n * create an eval from it\n * debug it,\n * or let Agnost fix it itself: our GitHub app opens a PR, tested against your past conversations, so you know it works before it ships.\n\n Your agent gets better every week.\n\n### **Who we are**\n\nChildhood friends, 8 years and counting.\n\n![uploaded image](/media/?type=post&id=107260&key=user_uploads/2055517/e3c2eaad-8b6e-4ec0-abfa-b16074688963)\n\n**Shubham**: Youngest engineer on Cisco’s analytics team, First Hire at Formbricks, Wrote code in Bitcoin, OWASP, and the Linux Foundation.\n\n**Parth**: IIT Madras CS (ranked 159 of 1.5M), built Spark pipelines at Microsoft, Led GPU tech at Infurnia (\\~$1M ARR).\n\n### **Our Asks**\n\n1. **Building a chat or voice agent?** Sign up & I’ll personally find 3 insane things going on in your agent traces that you’d have no idea about.\n2. **Know a founder whose product is an agent?** Intro us, 40’ monitors (to monitor your agents) on us: [shubham@agnost.ai](mailto:shubham@agnost.ai)\n3. **Try it:** Free tier:[ agnost.ai](http://agnost.ai)", "url": "https://www.ycombinator.com/launches/Ru0-agnost-ai-product-analytics-for-ai-agents", "votes": 191, "created_at": "2026-07-24T15:53:30.811Z" } ] }, { "id": 33020, "name": "Radley", "slug": "radley", "former_names": [ "Radley AI, Inc" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/eaa9729cc14d1a256b46037b05b50ac5b7f7c291.png", "website": "https://radley.health/", "all_locations": "San Francisco, CA, USA; Remote", "long_description": "Radley is the first AI-native radiology practice. We're acquiring a radiology practice bringing Radley to $4.6M in annualized revenue and rebuilding it with agents from the ground up.\r\n\r\nRadiology is already in crisis: there aren't enough radiologists to interpret 1 in 6 medical scans. In 10 years, that gap is growing to 1 in 3 - 246M scans a year.\r\n\r\nFrom past experience, we know that point solutions won’t cut it. While there is demand (we have $2.2M in contracts to prove it), bottleneck isn't just software, it's outdated systems and slow, inefficient operations. That's why we own the full stack, clinical and operational, end to end.\r\n\r\nWe met at Carbon Health as we scaled operations 31x and expanded to 100+ clinics. Now we're rebuilding radiology.", "one_liner": "The AI-native radiology practice", "team_size": 2, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1786061168, "tags": [ "B2B", "Healthcare" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada", "Remote", "Fully Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/radley", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/radley.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/radley-ai/", "x_url": "https://x.com/RadleyAI", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/eaa9729cc14d1a256b46037b05b50ac5b7f7c291.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 1524974, "full_name": "Maryam Jahed", "title": "Founder/CEO", "founder_bio": "CEO Founder at Radley. Previously Director of Product at Carbon Health (31x growth to 100+ clinics in 2 years). Head of Product at Midday (exit to Fortune 100 insurer). Co-founded Airo Health (YC-backed mental health AI; six-figure revenue). Built FDA breakthrough-designated CBT chatbot at Woebot. BASc Electrical Engineering, BA Sociology from Waterloo.", "is_active": true, "linkedin_url": "https://linkedin.com/in/maryamjahed", "x_url": "https://twitter.com/MaryamJM", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/63808a4521ef9ee494fda95f2dbe77c9ec47c069.jpg" }, { "user_id": 3729910, "full_name": "Seak Meng Lay", "title": "Founder/CTO", "founder_bio": "CTO Founder at Radley. Founding engineer and later VP of Software Engineering at Carbon Health, where I led engineering as the company scaled 31× to 100+ clinics serving millions of patients. Previously an engineer at Twitter. MIT CS.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/seakmenglay/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/41d7bad8ebca0d01b9a9e969095e394183c563d3.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111044, "title": "Radley: We're building the first full stack, AI-native radiology practice", "tagline": "We're redesigning the entire radiology operation to be AI-native from the ground up, unlocking scan capacity without adding radiologists. ", "body": "**TL;DR: We're Corgi for Radiology. The first AI-native full stack radiology practice. We’re acquiring a radiology practice bringing us to $4.6M in annualized revenue.** \n\n**Problem.** Radiology is already in crisis. The US needs 20% more radiologists to meet demand, and in 10 years that gap will reach 50% - 246M scans the workforce can’t handle. Agents can finally do most workflows. The industry is not moving fast enough to adopt it. \n\n**How we started.** Radiology is flooded with AI hype. We started riding that wave. But after digging into how these practices actually run, and drawing on our time scaling Carbon Health to 100+ sites, one thing became obvious: selling a dozen AI point-solutions and rolling them out department by department is too slow, and only adds incremental value. Building the whole practice in a AI-native way from day one is what’s needed. It solves the radiologist shortage directly instead of one contract at a time. This shift grew our market **21x**.\n\nSame insight Corgi had in insurance: don't sell software to the incumbent, become the AI-native incumbent.\n\nSo now we’re building the entire radiology workflow and business operations AI native and adding capacity without headcount. \n\n\n\n**Traction:**\n\n* **$4.6M annualized revenue**\n\n**Why us.** We met at Carbon Health as we scaled operations 31x and expanded to 100+ clinics. Now we're rebuilding radiology.\n\n**Our ask. Reach out if:**\n\n* You’re a **radiologist** who wants to work at and shape an AI-native practice\n* You can help connect us with an independent or hospital owned **imaging centers** struggling to keep up with radiology demand \n* Anyone who's building in **AI-native services**, we'd love to compare notes\n\nReach us at [founders@radley.health](mailto:founders@radley.health).", "url": "https://www.ycombinator.com/launches/St2-radley-we-re-building-the-first-full-stack-ai-native-radiology-practice", "votes": 41, "created_at": "2026-08-18T16:30:13.266Z" } ] }, { "id": 33021, "name": "Care GP", "slug": "care-gp", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/3f871fa8c32b501e71ef06d881893d564e003fbd.png", "website": "https://caregp.com.au/", "all_locations": "Sydney, NSW, Australia", "long_description": "Care GP is building AI agents to run operations for primary healthcare clinics", "one_liner": "AI agents to run primary healthcare operations", "team_size": 10, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1780569041, "tags": [ "SaaS", "Healthcare", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "Australia", "Oceania" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/care-gp", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/care-gp.json", "year_founded": 2024, "city": "Sydney", "country": "AU", "linkedin_url": "https://www.linkedin.com/company/caregp", "x_url": "", "fb_url": "https://www.facebook.com/profile.php?id=61576741965683", "cb_url": null, "github_url": "https://github.com/orgs/Care-AI-Inc", "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/3f871fa8c32b501e71ef06d881893d564e003fbd.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 1162884, "full_name": "Melvin Chen", "title": "Founder/CEO", "founder_bio": "Founder & CEO at Care GP (YC S26). \n\nPreviously Co-founded Avigo and StarLake Group. Held roles in private equity at Terra Firma, Samena Capital and Goldman Sachs.\n\nGraduated Valedictorian with an Engineering triple major in Materials, Economics and Management from the University of Oxford. Featured in the Forbes ‘30 Under 30’ list for Finance & VC. Born in China and raised in the UK.\n\nLoves sports and fitness - lately the racket sports and golf if anyone wants a hit!", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/melvinchen/", "x_url": "https://x.com/mchenofficial", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/da150cc9e81f4804c00640fe14ca77d86fbdca5c.jpg" } ], "jobs": [], "news": [], "launches": [] }, { "id": 33022, "name": "Praxis Robotics", "slug": "praxis-robotics", "former_names": [ "Praxis AI" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/9bc18b804ad9a5d92ef1db88e90491f98e0204a7.png", "website": "https://www.praxisrobotics.io", "all_locations": "San Francisco, CA, USA", "long_description": "Praxis captures the data physical AI is bottlenecked on: egocentric video, 3D scans, and multimodal capture from inside real industrial and residential environments. Embedded within publicly listed and unicorn-scale conglomerates, we reach 60k workers across 5 continents and 150+ environment types, supplying the real-world training data frontier labs and humanoid companies can't source anywhere else.", "one_liner": "Turns real businesses into robotics data vendors", "team_size": 3, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1782948528, "tags": [ "Hard Tech", "Reinforcement Learning", "Robotic Process Automation" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/praxis-robotics", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/praxis-robotics.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/130274179/admin/dashboard/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/9bc18b804ad9a5d92ef1db88e90491f98e0204a7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 2794157, "full_name": "Rohan Seelamsetty", "title": "Founder/CEO", "founder_bio": " CEO @ Praxis. Forged across factory floors, campaign trails, and niche forums. Ex Computation and Cognition at Penn, Apollo Fellow at Oxford, and Indian Debate Team (#1 ranked team worldwide)", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rohan-s-1186b2214", "x_url": "https://x.com/ro_seelamsetty", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4b3e6e16cddb3da2384ce8f9439ca9ebc043d2bb.jpg" }, { "user_id": 3752185, "full_name": "Dev Karpe", "title": "Founder/COO", "founder_bio": "Founder at Praxis AI, previously at Wharton's Huntsman Program. Background in investment banking and supply chain ops at a F500 CEO's office. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dev-karpe", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f4a1fc8056a887a99633492428b5db995e58a488.jpg" }, { "user_id": 3748464, "full_name": "Tommy Li", "title": "Founder/CTO", "founder_bio": "Founder at Praxis AI. Prev. tech private equity at Morgan Stanley. Industrial tech venture capital at Andreessen Horowitz. Published IEEE researcher at NASA JPL and PerceptIn Robotics. BSE CS and Finance from UPenn.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/tommyli-penn/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/0e7be9454dacfd8fa65594cc6836b90cc7e5e0b4.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109471, "title": "Praxis AI - Training Data for Robotics", "tagline": "Praxis offers labs the most diverse environments, SOTA hardware, and rigorous post processing to train their models", "body": "**Hey YC!**\n\nWe're Rohan, Dev, and Tommy, and we're building Praxis.\n\n**Praxis turns real-world work into training data for robots.**\n\n\n\n**What we do**\n\nRobots learn to act in the physical world by watching humans do the work. We partner with businesses to capture egocentric human demonstration data, the first-person video of how tasks actually get done in factories, warehouses, homes, and other physical work environments, then supply it to the labs and humanoid teams training the next generation of robots. Robotics data is our focus, but since our collection infrastructure overlaps, we're happy to explore other multimodal data too.\n\n**The problem**\n\nRobotics is missing its internet-scale training corpus. LLMs learned from the web, books, code, and the accumulated record of human thought. Robots need the equivalent record of human action: how people pick, pack, clean, assemble, inspect, cook, use tools, recover from mistakes, and move through messy physical environments.\n\nCollecting some of this data is easy. Collecting enough of the right kind is brutally hard, for two reasons.\n\nFirst, diversity is capped by access. Models only generalise if they've seen the full range of real environments and operators, from a one-person shop to a national-scale industrial operation. Getting in the door across that entire spectrum, and across continents, is a physical-world access problem that most teams can't solve.\n\nSecond, it takes two rare skills at once. You have to know what frontier labs need at a technical level, and be able to run demanding physical operations across thousands of real sites, standardising capture and turning fragmented workflows into model-ready data. Few teams have both halves.\n\n**Our solution**\n\nWe already run collections across 150 environment types on 5 continents, with diversity from mom-and-pop stores to conglomerates accounting for 3% of a nation's GDP, and a reachable network of 60,000 workers. Depending on the task, we capture egocentric video, stereo-RGB, IMU, wrist-camera views, glove kinematics, haptics, and dense annotations. Everything is delivered calibrated, synced, pose-tracked, QA'd, and PII-scrubbed. We build and deploy custom capture hardware quickly because part of the team operates out of Shenzhen.\n\n**Why us**\n\nPraxis is built to do both hard things at once. We pair robotics research depth (IEEE, NASA JPL) with the ability to actually run demanding operations in the real world, from Fortune 500 supply chains to Parliaments, plus a Shenzhen hardware team. That combination- technical understanding of what labs need plus real physical-world access at scale- is what lets us deliver diverse, model-ready data at a cost and breadth that is irreplaceable.\n\n**Our asks**\n\n1. Intros to anyone working on robotics, embodied AI, or world-models, especially data and research leads. If you need other forms of multimodal data, we're happy to explore that too.\n2. If you run a business with interesting physical work (logistics, manufacturing, food, retail, or lab work) and would like to monetize it, we'd love to talk.\n\nYou can reach us at [founders@praxisrobotics.io](mailto:founders@praxisrobotics.io) and [rohan@praxisrobotics.io](mailto:rohan@praxisrobotics.io)[ https://www.praxisrobotics.io](https://www.praxisrobotics.io)", "url": "https://www.ycombinator.com/launches/STf-praxis-ai-training-data-for-robotics", "votes": 11, "created_at": "2026-08-06T22:30:00.618Z" } ] }, { "id": 33023, "name": "Levocred AI", "slug": "levocred-ai", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/4211252c17df7b0a89dc9c36ce49996bef59474d.png", "website": "https://levocred.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Levocred AI build AI employees for lenders including credit funds, fintechs, credit unions & banks to automate operations like servicing/collections, underwriting, back-office tasks etc. ", "one_liner": "AI Workforce for lenders", "team_size": 2, "industry": "Fintech", "subindustry": "Fintech", "launched_at": 1782022398, "tags": [ "Finance", "B2B", "Enterprise Software" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/levocred-ai", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/levocred-ai.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/levocred-ai", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/4211252c17df7b0a89dc9c36ce49996bef59474d.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3179062, "full_name": "Mohit Gupta", "title": "Founder/CEO", "founder_bio": "Employee #1 at a credit fund that went from $10M to $1B AUM. Managed $200M at 23, evaluated $100B of loans. Ex Morgan Stanley fixed income. CS @ IIT Bombay, Finance @ UC Berkeley. Now building the Bloomberg of loans.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/guptamohit2299/", "x_url": "https://x.com/gupta_m0hit", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3ec48e4ead1bd6681f312cacd75cb1ddb691cd74.jpg" }, { "user_id": 1978755, "full_name": "Saksham Gupta", "title": "Founder/CTO", "founder_bio": "Founding engineer at a credit fund, where he built and scaled the trading and data infrastructure as the fund grew from $150M to $1B+ in AUM. EE @ IIT Kanpur", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/gupsak", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d005e1efc44e74b6946d4413ff8c73a055def3b8.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109810, "title": "Levocred: AI platform for credit funds, fintechs and banks", "tagline": "AI for credit funds, fintechs and banks", "body": "---\n\n**TLDR:** Levocred AI (YC S26) is an AI-native platform for credit funds, fintechs and banks to automate workflows like back-office tasks, portfolio monitoring, compliance and underwriting. Live across multiple credit teams, running \\~$1Bn in receivables through the platform.\n\n\\*\\*Video:\\*\\*\\\n\n\nThe winning teams will scale with token consumption, not headcount. Try it today: \n\n**The problem**\n\nCredit teams run on data scattered across 5-6 tools (CRM, LMS, data rooms, servicer reports) and endless spreadsheets. So highly-paid analysts become data janitors, copy-pasting between systems and rebuilding the same models — instead of the real job: underwriting and managing risk.\n\nIt's the same story across all of an analyst’s workflows (like new deals, monitoring, reporting, workouts). One example: to borrow against a credit facility (it’s like a credit card for fintechs with a limit typically in the $100Ms ), a team produces and verifies compliance certificates by hand. That's hours of Excel every time, so they draw less often than they should, leaving millions on the table.\n\n**What is Levocred?**\n\nLevocred is an AI employee for your credit team. It plugs into your loan tapes, credit agreements, servicer reports, and bank feeds, learns how you work, and runs the workflows your analysts do by hand today.\n\n* Analyzes new deals and drafts the investment memos\n* Monitors the portfolio, marks valuations, and answers questions in Slack (\"what's our exposure to X?\")\n* Produces and checks compliance certificates, and handles covenant and LP reporting\n\nWhy it works when nothing else has:\n\n* **It's AI-native, not a rules engine.** Older tools only work if you hand-code every rule, so they break on messy, real-world data. Levocred's agents read your actual documents and tapes and reason across them. We already run consumer, fintech, litigation finance, and MCA through the same engine.\n* **You stay in control** with full auditability and no waiting on a vendor's engineers\n* **It compounds.** Every document you load makes it smarter about your book. That data moat is yours. It’s why usage only grows.\n\nThe result: a draw that took a day of Excel takes minutes, monitoring runs continuously instead of monthly, and your team spends its time on judgment, not grunt work.\n\n**Team**\n\n* **Mohit Gupta (CEO)**: Studied CS at IIT Bombay (ranked 53rd of 1.2M in the JEE) and Finance at UC Berkeley. Founding engineer at a credit hedge fund, where he later managed a $200M book.\n* **Saksham (CTO)**: Previously Software Engineer at a credit fund, where he built and scaled infrastructure as the fund grew. Electrical & Electronics from IIT Kanpur", "url": "https://www.ycombinator.com/launches/SZ8-levocred-ai-platform-for-credit-funds-fintechs-and-banks", "votes": 5, "created_at": "2026-08-11T15:30:00.232Z" } ] }, { "id": 33045, "name": "Axelrod", "slug": "axelrod", "former_names": [ "Axelrod Labs Inc", "Axelrod Labs" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/7fbd96bca4da9de9db65dd46c8e36b9fdaa3e168.png", "website": "https://axelrodlabs.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Autonomous software for hotels that takes over all front-of-house and back-of-house tasks with zero integrations, running operations better than your staff would. Axelrod is trusted by Michelin Key awarded and Marriott hotels and is live across 3 continents.", "one_liner": "Hotels that run themselves", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1784236014, "tags": [ "Travel", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/axelrod", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/axelrod.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/axelrod-labs/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/7fbd96bca4da9de9db65dd46c8e36b9fdaa3e168.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3649756, "full_name": "Adrian Lucas Tariel Stoica", "title": "Founder", "founder_bio": "Founder at Axelrod (YC S26). Previously research on AI Awareness. Co-produced records totaling 40M+ streams at Sony Music. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/stoicaadrian/", "x_url": "https://x.com/AdrianLTStoica", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4659c00be13b9fc81aaa556a654964dc49ceb191.jpg" }, { "user_id": 2112304, "full_name": "Saman Sayahpour", "title": "Founder", "founder_bio": "Founder at Axelrod (YC S26). Dropped out of Carnegie Mellon's B.S. in Logic & Computation. Previously researched causal machine learning, prediction markets, photonic computing and collective decision-making. Admitted to study mathematics at the University of Vienna at 14.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/sayahpour/", "x_url": "https://www.x.com/samansayahpour", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/0a5faa6a842709cc020329e50bbf601398fd65f0.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109442, "title": "Axelrod - Autonomous agents that run luxury hotels", "tagline": "Agents that execute every non-physical front- and back-of-house task inside the hotel's existing system", "body": "**TL;DR:** Axelrod deploys agents in boutique and luxury hotels. Agents execute all non-physical front-of-house (FoH) and back-of-house (BoH) tasks inside the hotel's existing systems, just like your staff would, in order to deliver a hyper-personalized guest experience.\n\n\\\n**The problem**\n\nRunning a hotel is a grind of manual, repetitive work. Staff are occupied with menial administrative tasks that do not directly improve the guest experience. Small operators burn their own hours on it, large operators hire teams. Either way it kills time or margins. Existing software is just one more thing to manage, it doesn't do the work for you.\n\n\n\n**How we solve it**\n\nAxelrod maps the hotel's entire operation, then runs an agent workforce on top of it. The harness is purpose-built for hotel operations, powered by a memory layer that holds your property’s standard operating procedures and every guest’s history across stays, as well as an orchestration layer that breaks a request into tasks, routes each to the right role-agent and coordinates them across departments.\n\nWith computer- and browser-use, our agents drive the hotel's existing stack (PMS, POS, RMS) the way your staff do, with no MCPs, APIs, or integrations. That’s how Axelrod runs on the fragmented, legacy software real hotels actually use.\n\nAxelrod augments every single role in the hotel staff: one butler can cover 4x the rooms. Because each agent has full context, it can adequately assess guest needs and makes sure nothing falls through the cracks. Agents start fully supervised and earn autonomy as they prove themselves, so hotels hand over the operations at their own pace. \n\nWe’re trusted by Michelin-Key awarded hotels and live across four continents.\n\n**Why us**\n\nWe both grew up in the hospitality industry and have worked on both sides of the desk.\n\n**Our ask**\n\nIntros to executives at hotel management groups, luxury hotel owners and general managers as well as hotel CTOs/CIOs.\n\n**Contacts**\n\n[Adrian Lucas Tariel Stoica](https://www.linkedin.com/in/stoicaadrian/): \\\n[adrian@axelrod.live](mailto:adrian@axelrod.live) \\\n\\\n[Saman Sayahpour](https://www.linkedin.com/in/sayahpour/): \\\n[saman@axelrod.live](mailto:saman@axelrod.live)", "url": "https://www.ycombinator.com/launches/STC-axelrod-autonomous-agents-that-run-luxury-hotels", "votes": 13, "created_at": "2026-08-06T13:00:12.647Z" } ] }, { "id": 33046, "name": "CarSignal", "slug": "carsignal", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/adebdc425b676af3d7f53e7ba2ba70bd7359f2f7.png", "website": "https://www.trycarsignal.com", "all_locations": "San Francisco, CA, USA", "long_description": "CarSignal is the AI-native operating system for independent auto repair shops.\r\n\r\nMost shops still run each repair across disconnected management software, scan tools, parts catalogs, service information, texts, spreadsheets, and knowledge kept in technicians’ heads. That fragmentation slows diagnosis, creates duplicate processes, and leaves customers waiting.\r\n\r\nCarSignal connects the full repair process in one system. AI agents research vehicle-specific failure patterns and surface the tests that confirm or rule out each cause. The same system carries each job from booking and intake through inspections, estimates, parts, approvals, payment, and follow-up, so technicians spend less time searching and updating systems and more time fixing cars.", "one_liner": "The AI Operating System for Auto Shops.", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1785653742, "tags": [ "Hardware", "SaaS", "B2B", "AI", "Automotive" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/carsignal", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/carsignal.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/adebdc425b676af3d7f53e7ba2ba70bd7359f2f7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3650329, "full_name": "Michael Muzzin", "title": "Founder/CEO", "founder_bio": "Founder at CarSignal (YC S26). Previously electrical engineering at NVIDIA, he authored the Factory Failure Disposition Guide governing re-qualification of $300M in HGX and DGX datacenter hardware. BSEE, UCLA.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/mmuzzin/", "x_url": "https://x.com/muzzin_michael", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/916db4a926a764d2aa5180ad114eb41c9395345a.jpg" }, { "user_id": 3485129, "full_name": "Junaid Popalzai", "title": "Founder/CTO", "founder_bio": "founder at carsignal. prev netflix, notion, whatnot, tesla, amazon. eecs @ berkeley", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/junaid-popalzai-8b3753273", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8810a567ae33eca8713595d869e17c62401df9ce.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110407, "title": "CarSignal: The AI Operating System for Auto Shops", "tagline": "Your shop runs on one AI platform: booking, diagnostics, quoting, parts, and payments all in a single loop.", "body": "**TL;DR**:\\\nCarSignal is the AI Operating System for Auto shops. CarSignal provides a complete end to end operating system for mechanics and shop owners. We’re live in over 40 shops across 18 US states and are expanding everyday. \n\n\n\n🤯 **The Problem**\n\nThe average auto repair shop in the US runs across multiple different systems, none of which talk to each other. With over 300,000 of these shops existing in the US, the problem is vast. Owners are trapped paying for a web of complex disjointed software in order to maintain basic functionality. \n\n🛠️ **The Solution**\n\nCarSignal fixes those discontinuities, streamlining recurring tasks that every shop runs daily. We deploy four Agents across the traditional auto shop stack to boost efficiency and lower lift time. A voice agent is deployed as the shop's receptionist, a custom OBD-II scanner provides factory level context to a diagnostic agent who aids in finding root cause issues in the vehicle. CarSignal also provides an estimate agent who finds real parts and labor times across wholesalers to give accurate estimates instantly, and a marketing agent who keeps track and reminds customers of overdue service and scheduled maintenance.\n\n👋 **Team**\\\nMichael Muzzin - UCLA Electrical Engineering, ex-Nvidia\n\nJunaid Popalzai - UC Berkeley EECS, ex- {Tesla, Amazon, Notion, Netflix}\n\n**Our Ask**\\\nIf you know a mechanic, or are a mechanic, book a demo! We’d love to show you what CarSignal is capable of.\n\ntrycarsignal.com/demo", "url": "https://www.ycombinator.com/launches/Sil-carsignal-the-ai-operating-system-for-auto-shops", "votes": 24, "created_at": "2026-08-17T16:30:09.423Z" } ] }, { "id": 33047, "name": "Decawork", "slug": "decawork", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/44fee5cae005d704af84e7ed5ed7418d4ca0580c.png", "website": "https://decawork.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "An agent’s access should depend on the task it is doing. The same action can be appropriate for one task and wrong for another. Access decisions need to adapt in real time as the agent works, but companies can’t turn every business judgment into a fixed rules or have people manually approve every step.\r\n\r\nWe train small, company-specific models that approve AI actions in real time. They consider the company’s policies and interests, the agent’s current task, and its past actions to decide whether each proposed action should be allowed. Access is specific to the task at hand and reassessed as the agent works.\r\n\r\nCompanies will rent intelligence from AI labs or run open-source models to power their agents. They’ll own the models that keep their entire AI workforce acting in their interests. Decawork is building those models.", "one_liner": "Training company-specific models that approve AI actions.", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1785265630, "tags": [ "B2B", "Security", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/decawork", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/decawork.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/decawork-ai/", "x_url": "https://twitter.com/decawork", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/44fee5cae005d704af84e7ed5ed7418d4ca0580c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/bea64bca8f89d0506e06952dbd00022b772363d1.png" ], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 1632535, "full_name": "Aman Raj", "title": "Founder/CEO", "founder_bio": "- Built AI compliance platform at Barclays used by internal teams globally\n- Founded a fintech and scaled it to 52K families and $1.5M AUM\n- Led product at a consumer AI startup, scaled to 100K+ MAU (Google Play Best App 2024)\n- IIT Kharagpur", "is_active": true, "linkedin_url": "https://linkedin.com/in/aman-raj-84636a158", "x_url": "https://x.com/arxneo14", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2ae56b87789485359a28cd015856039060019efe.jpg" }, { "user_id": 1049492, "full_name": "Sarthak Aggarwal", "title": "Founder/CTO", "founder_bio": "- AI systems at NVIDIA (deployed at OpenAI, Meta)\n- Enterprise AI at Ema (used by Microsoft, Hitachi)\n- Led Conquest, Asia's largest student-run accelerator\n- Google Code-In Global Winner + Open source since the age of 12\n- BITS Pilani", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/sarthak-agg/", "x_url": "https://x.com/_sarthak4", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f93e2ff857332071c545c3c16f569835408cf579.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109883, "title": "Decawork - Training company-specific models that approve every AI action", "tagline": "Our small AI models decide agent access in real time, task by task, based on your company’s interests.", "body": "Hi everyone 👋 - we're Aman and Sarthak, founders of Decawork.\\\n\\\n**TL;DR:** AI agents shouldn’t have permanent access to company systems. The same action can be appropriate for one task and wrong for another, but companies can’t turn every business judgment into fixed rules or manually approve every step. Decawork trains small, company-specific AI models that approve every agent action in real time, based on company policies and interests, the agent’s task, and its past actions. Access is granted for the task at hand and revoked when it’s done.\\\n\\\nLaunch video: [Watch it here](https://www.youtube.com/watch?v=SanpHIyFcV4)\\\n\\\n🚧 **The Problem & Why now?**\\\nCompanies are putting long-running, more autonomous agents to work on broader tasks and goals. These agents need access to critical company systems, but their actions can be unpredictable and may not serve the company’s interests.\n\nOpenAI’s agents have already hacked Hugging Face during testing with reduced safeguards. Replit’s agent deleted data from a live application’s database.\n\nAn agent’s access should depend on the task it is doing. The same action can be appropriate for one task and wrong for another. Permanent access lets an agent keep using permissions even when they are no longer appropriate for its task.\n\nAs agents take on more work independently, access decisions need to happen in real time. But companies can’t turn every business judgment into a fixed rule or have people manually approve every step.\\\n\\\n✅ **How we solve this?**\\\nDecawork gives AI agents access for the task at hand and revokes it when the task is done.\n\n* Connect your agents. Route API requests and MCP tool calls through Decawork to see activity and manage access in one place.\n* Set company policies. Write policies in plain English, specify when human approval is required, and connect your existing identity and IT service management systems.\n* Deploy a company-specific policy model. We deploy small models that approve actions in real time using your company’s policies and priorities, task context, and past actions. Your agents do the work. Decawork evaluates each proposed action and enforces the access decision before it runs.\n\n🚀 **Team**\\\nSarthak built AI systems at NVIDIA that were deployed at OpenAI and Meta, and worked on enterprise AI agents at Ema.\n\nAman built the AI compliance platform at Barclays used by internal teams globally, then led product at a consumer AI startup serving 100K+ monthly users, hand-wiring credentials for the agents he shipped.\n\nEvery one of those deployments came down to the same questions: what can it access, who approved it, and what did it do? We kept solving agent access by hand - so we quit to build the system we wished we’d had.\n\n![uploaded image](/media/?type=post&id=109883&key=user_uploads/1632535/0ef2784e-ab45-422f-91b9-960473f11dd5)\n\n🙏 **Our Ask**\n\n* If you’re on an IT team or know someone who’s worried about AI agents accessing or misusing company systems, talk to us: or [founders@decawork.ai](mailto:founders@decawork.ai)", "url": "https://www.ycombinator.com/launches/SaJ-decawork-training-company-specific-models-that-approve-every-ai-action", "votes": 152, "created_at": "2026-08-12T16:30:00.369Z" } ] }, { "id": 33048, "name": "Pennant", "slug": "pennant", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/7fbb4be6497bcd38a1210a0561cf3eb84422ab32.png", "website": "https://www.getpennant.ai", "all_locations": "New York City, NY, USA", "long_description": "Led by former institutional proxy voters and technologists with a collective 20+ years of experience, Pennant gives institutional investors, public companies, and corporate advisors voting and governance intelligence to act on their own principles at scale.", "one_liner": "The World Model for Corporate Governance", "team_size": 2, "industry": "Fintech", "subindustry": "Fintech -> Asset Management", "launched_at": 1786584832, "tags": [ "Fintech", "Investing", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech", "Asset Management" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/pennant", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/pennant.json", "year_founded": 2026, "city": "New York City", "country": "US", "linkedin_url": "https://www.linkedin.com/company/pennant-ai", "x_url": "https://x.com/pennant_ai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/7fbb4be6497bcd38a1210a0561cf3eb84422ab32.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/642317534b66a99aea64b59d023b97471b2e61fe.png", "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/5dd4ad2aebecc8b69db01a98a2fc2aa9ffedb2e8.png" ], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3388610, "full_name": "Ryan Nowicki Stewart", "title": "Founder/CEO", "founder_bio": "Co-Founder at Pennant. 10+ years asset management, proxy voting & shareholder activism industry experience across PJT, State Street and BlackRock. Palantir Startup Fellowship Cohort 002. Cornell U. Economic Club of NY. Hoover Institution.\n\n\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ryan-nowicki-stewart/", "x_url": "https://x.com/rywicks", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2ce9a1682e6222638bce37b0d89bdf0e78b7aeb8.jpg" }, { "user_id": 473596, "full_name": "Tomas Taylor", "title": "Founder/CTO", "founder_bio": "Co-Founder at Pennant. Previously PM at BlackRock. Founded Mindware, an agent tooling infrastructure startup, before MCP. Also co-founded ClawCon, the fastest-growing community for Personal AI.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/tomastaylor", "x_url": "https://x.com/0xtotaylor", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/cf9b1177e1c0378455a6b6503c630ed7aed90706.jpg" } ], "jobs": [ { "id": 111722, "title": "Founder Fellow - GTM, Growth & Strategy", "url": "https://www.ycombinator.com/companies/pennant/jobs/21lflHF-founder-fellow-gtm-growth-strategy", "location": "New York, NY, US", "type": "Internship", "role": "Sales", "role_type": null, "salary_range": "", "equity_range": "", "experience": null, "visa": "US citizen/visa only", "skills": [ "Data Analytics", "Figma", "Evals", "AI Agents", "AI Safety", "Generative AI" ] }, { "id": 111727, "title": "Founder Fellow - Applied AI", "url": "https://www.ycombinator.com/companies/pennant/jobs/eUFGOW1-founder-fellow-applied-ai", "location": "New York, NY, US", "type": "Internship", "role": "Engineering", "role_type": "AI engineer", "salary_range": "", "equity_range": "", "experience": null, "visa": "US citizen/visa only", "skills": [ "Google Cloud", "Node.js", "PostgreSQL", "Python", "TypeScript", "SQL", "Prompt Engineering", "LLMs", "Evals", "AI Agents", "BigQuery", "Generative AI", "OpenAI", "LangSmith" ] }, { "id": 111752, "title": "Founding Engineer, Data & AI", "url": "https://www.ycombinator.com/companies/pennant/jobs/z6mIB9W-founding-engineer-data-ai", "location": "New York, NY, US", "type": "Full-time", "role": "Engineering", "role_type": "Backend", "salary_range": "$150K - $250K", "equity_range": "0.25% - 1.50%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [ "Go", "Google Cloud", "Kubernetes", "Node.js", "PostgreSQL", "Python", "TypeScript", "SQL", "Docker", "ETL", "gRPC", "Terraform", "LLMs", "AI Agents", "LangChain", "BigQuery", "Generative AI", "NestJS", "OpenAI", "LangSmith", "Datadog" ] }, { "id": 111753, "title": "Founding Engineer, Product", "url": "https://www.ycombinator.com/companies/pennant/jobs/Y5USzcY-founding-engineer-product", "location": "New York, NY, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$150K - $250K", "equity_range": "0.25% - 1.50%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [ "Google Cloud", "Node.js", "PostgreSQL", "React", "TypeScript", "SQL", "Next.js", "LLMs", "AI Agents", "Generative AI", "NestJS", "Vercel", "Playwright", "OpenAI", "LangSmith", "Datadog" ] } ], "news": [ { "title": "EP28: Proxy Voting Is Always On: Pennant’s Ryan Nowicki Stewart on AI-Native Proxy Voting", "url": "https://www.thefundaipod.com/episodes/episode/33ed71e2/ep28-proxy-voting-is-always-on-pennants-ryan-nowicki-stewart-on-ai-native-proxy-voting", "date": "Oct 13, 2026" }, { "title": "AI tools emerging in proxy voting - Pensions & Investments", "url": "https://www.pionline.com/technology-innovation/artificial-intelligence/pi-ai-tools-emerging-proxy-voting/", "date": "Oct 04, 2026" }, { "title": "START: Ryan Nowicki Stewart, CEO & Co-Founder, Pennant: \"The Corporate Governance OS for Public Markets\" | Fondo", "url": "https://fondo.com/blog/start-ryan-nowicki-stewart?utm_term=ryan-nowicki-stewart-pennant&utm_campaign=10061329-Startup%20Growth%20Podcast&utm_content=386749028&utm_medium=social&utm_source=twitter&hss_channel=tw-1843505047692554240", "date": "Sep 30, 2026" }, { "title": "Boardroom Governance - Ryan Nowicki Stewart", "url": "https://podcasts.apple.com/in/podcast/ryan-nowicki-stewart-ceo-pennant-ai-ai-proxy-voting/id1513064579?i=1000792196213", "date": "Sep 29, 2026" }, { "title": "Cornell University: Successful Alumni Who Founded Startups - TechRound", "url": "https://techround.co.uk/startups/cornell-university-successful-alumni-who-founded-startups/", "date": "Sep 25, 2026" } ], "launches": [ { "id": 110156, "title": "Pennant - The Corporate Governance OS for Public Markets", "tagline": "AI-native proxy voting and governance intelligence, built by the team who used to do it by hand.", "body": "**TL;DR:** Pennant is building AI-native proxy voting solutions for institutional investors, public companies, and their advisors. We’re live with paying enterprise customers, including one of the largest US public pension plans, a Top AmLaw 100 firm, and leading governance advisors. Before Pennant, our team built trillion-dollar tech at BlackRock, cast proxy votes at State Street, and advised Fortune 100 boards at PJT.\\\n\\\n[https://drive.google.com/file/d/1_WDs1VB213EcPS5caE5XK7L6UkZLz07l](https://drive.google.com/file/d/1_WDs1VB213EcPS5caE5XK7L6UkZLz07l/preview)\n\n**The problem:** Proxy voting at publicly traded companies drives trillions in market value every year. Shareholder votes decide $100B mergers, trillion-dollar pay packages, and the boardroom battles that define companies.\n\nYet the work is brutally manual: reviewing thousands of pages of SEC filings, research, and engagement notes that live in a patchwork of PDFs, spreadsheets, and post-it notes. And for professional advice, the market has relied on two firms, ISS and Glass Lewis, for 40+ years, a duopoly that controls 97% of the market.\n\n**Why Now:** Technological change and market sentiment are ripping this market wide open.\n\nCEOs have been sounding the proxy-advisor alarm for years:\n\n![uploaded image](/media/?type=post&id=110156&key=user_uploads/3388610/45604eff-9d48-4271-90ed-b3bf6000b461)\n\n![uploaded image](/media/?type=post&id=110156&key=user_uploads/3388610/a5232d8f-91b7-44bd-8116-f0d251b764d0)\n\nAnd now big investors and the White House are taking action.\n\n![uploaded image](/media/?type=post&id=110156&key=user_uploads/3388610/12da7f65-6e5c-4e57-b2d5-76c18fac1fcf)\n\n![uploaded image](/media/?type=post&id=110156&key=user_uploads/3388610/00e9661e-2b7e-414d-8c2c-f4d4f47b8b55)\n\n![uploaded image](/media/?type=post&id=110156&key=user_uploads/3388610/47c99681-6ca9-484b-ab33-102a4021cd92)\n\nJPMorgan is even building its own solution. But 99% of the investors, companies, and advisors we talk to don't want to own this themselves, or don't have the resources to. That's where Pennant comes in, and it's why we're winning business out of the gate.\\\n\\\n**Why Pennant is better:** Pennant isn't a better proxy advisor. We're building something bigger: the AI-native operating system for corporate governance, built by former proxy voters and technologists, a team that cast these votes and shipped the software at institutional scale.\n\nPennant gives institutional investors, and the advisors and companies they engage with, voting and governance intelligence to act on their own principles at scale. Legacy tools follow rigid rules; Pennant reasons through each ballot item, inbound email, and engagement question the way your team would, handling the clear calls and flagging the nuanced ones, keeping people in the driver's seat.\n\nAs one of our customers put it last week, “\\[incumbent xyz\\] can handle our routine stuff for another year, but we’re thrilled to use Pennant for the custom heavy-lifting that we’ve been wanting to do for a long time, and use you for everything from there.”\n\nPennant also serves the other side of the vote: public companies, law firms, and bankers who need to read their shareholder base and predict risk before the annual meeting. Same data, bigger market. The market at the intersection of investing and corporate governance is estimated at well over $200 billion, and it remains largely unserved by today's tools.\n\n**Traction:** Pennant is live with paying enterprise customers, including one of the largest US public pension plans, a Top AmLaw 100 firm, and leading governance advisors. We're partnered with a top global asset manager as a design partner and are actively engaged with S&P 500 companies. Before YC S26, Pennant was selected for Palantir's Startup Fellowship (Cohort 002).\n\n**Why us:** We've known each other for 15 years, since our undergrad days at Cornell, and started our careers on the same analyst desk at BlackRock building tech products for these exact customers. We have a rare pairing: direct industry experience as the customer, plus the engineering chops to ship fast and earn trust.\n\nAfter BlackRock, Ryan (CEO) was the customer for 6+ years, voting proxies at 3,000+ company meetings and advising Fortune 100 companies on how to win critical votes at PJT Partners. He used Glass Lewis and ISS every day.\n\nTommy (CTO) took the technical path, shipping products at BlackRock before earning his Harvard MBA and founding Mindware, where he built workflow-ready infrastructure for AI agents. He was building agentic AI before \"agents\" were hot, and Pennant is where his engineering background meets Ryan's proxy-voting expertise.\n\n**Our ask:** If you own or touch proxy voting as an investor, advisor, or public company, we want to talk: \n\nLearn more at [getpennant.ai](http://getpennant.ai)\n\n![uploaded image](/media/?type=post&id=110156&key=user_uploads/3388610/d8fefc8c-cf1c-4ef8-a175-f112644f9ca7)\n\n", "url": "https://www.ycombinator.com/launches/Sei-pennant-the-corporate-governance-os-for-public-markets", "votes": 19, "created_at": "2026-08-12T18:30:00.287Z" } ] }, { "id": 33050, "name": "Derya", "slug": "derya", "former_names": [ "Derya Maritime / Meridian Forwarding (forwarding office)" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/e0512b453e994207fdf56052eb221e687d29dc51.png", "website": "http://usederya.com", "all_locations": "San Francisco, CA, USA", "long_description": "Derya deploys supply chain automations for startups, SMEs, and enterprises that fulfil their entire supply chain function, from prospecting and procurement to distribution of goods.\r\n\r\nOur mission is to rebuild the supply chain from the ground up, turning one of the world’s hardest coordination problems into an AI native system that moves commodities by being integrated into both the supply chain infrastructure and the customer. We work across robotics, commodities, and food & beverages companies; procuring goods and managing various operation functions for them.", "one_liner": "AI Brokers for Industrial Trade", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Supply Chain and Logistics", "launched_at": 1784839627, "tags": [ "AIOps", "Logistics", "Supply Chain", "Procurement", "Automation" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Supply Chain and Logistics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/derya", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/derya.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/derya-maritime", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/e0512b453e994207fdf56052eb221e687d29dc51.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/734e64627fe82417c09de0dc89e3ab054b81f94f.png" ], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 2282645, "full_name": "Mert Turna", "title": "Co-Founder", "founder_bio": "Building agentic supply chain automations @ Derya (YC S26). CS @ UCLA, built a ship operations software used in 30+ vessels ($200M Fleet Value).\n\nPast experience:\n- Co-Founder, @ Hemut (YC P25), truck management operations.\n- Several VC and IB internships\n- UCLA Debate Team Ops & Highest Speaker in the team\n- National Debating Team", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/mert-turna/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/12fc51cbbdae084e45eb08501c53376a7c27c1b8.jpg" }, { "user_id": 1979081, "full_name": "Oguzhan Karaca", "title": "Cofounder", "founder_bio": "Building agentic supply chain automations @ Derya (YC S26). Law at Oxford, built a 6-figure ARR education business on AI autopilot.\n\nPast experience:\n- CEO @ Ivy Consultants, scaling to six-figure ARR\n- Legal @ Ceva/CMA CGM\n- Editor @ Oxford Law Journal\n- Treasurer @ Pembroke (Oxford) Student Union\n- National Debating Team", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/oguzhan-karaca/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1bc38d2953cef10b16a1977c66084ed9a3eea127.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109885, "title": "Derya: Your supply chain autopilot for mapping, planning, and execution.", "tagline": "We build agents that help hardware companies design supply chain processes and eliminate friction.", "body": "TL;DR: We map out all your procurement, logistics, assembly, and distribution operations and then execute them for hardware companies. If you build hardware, physical AI, robots, we'll set up the operations while you focus on designing your product and sales.\n\n**Ask:** If you are spending too much time sourcing components, managing manufacturers, or coordinating shipments, let's see how we can save time and make money for you. Reach us at [contact@usederya.com](mailto:contact@usederya.com) or book a meeting at[ https://calendar.app.google/vSabvozH51TCM9V18](https://calendar.app.google/vSabvozH51TCM9V18) \n\n \n\n**The Problem**\n\nRedundant labor and inefficiencies represent $1.3 trillion of the $13 trillion logistics market. You shouldn't be spending your Tuesday chasing a customs broker, comparing three carrier quotes you don't trust, and hoping your parts clear China customs before your build slips a week.\n\n**Solution**\n\nWe’re setting up your supply chain for you:\n\n* **Shipping & customs automation** for cheaper and risk-free freight and imports\n* **Procurement** support to autonomously order, ship, and track all your inputs\n* **Production** visibility and execution to get your products ready to ship, quick\n\n**Why us**\n\nWe already handled more than $15M in freight before starting to handle the whole supply chain for companies ranging from lab robotics to biochemical manufacturers.\n\nWe’ve been friends for almost 10 years and built our first business together in high school. Mert was the CTO of the trucking AI company Hemut (YC P25), UCLA CS major, and the son of a ship captain. He built software for 25+ ships, a $450M fleet running on his software. \n\nOguzhan is an Oxford law grad who worked at CMA CGM/Ceva (3rd-largest shipper), previously built a six-figure-ARR AI-native business, and grew up in a farming family, providing him exposure as both the carrier and the customer.", "url": "https://www.ycombinator.com/launches/SaL-derya-your-supply-chain-autopilot-for-mapping-planning-and-execution", "votes": 7, "created_at": "2026-08-10T19:30:00.323Z" } ] }, { "id": 33051, "name": "Tokenless", "slug": "tokenless", "former_names": [ "Tokenless", "Touchy" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/7f250ec6cc81104148d8bd5bafdd6b96bc77d023.png", "website": "https://usetokenless.com", "all_locations": "San Francisco, CA, USA", "long_description": "Tokenless is an applied AI research lab. Our goal is to end reward hacking. We do research and build tools to detect issues with environments and evals, understand model behavior in those environments, and prevent those issues from happening. Correcting them makes training more effective, makes model performance easier to interpret, and is crucial to solving the AI alignment problem.\r\n\r\nIf you’re training models with RL, building environments, or running agent evals, and want more confidence in your environments, we're here to help.", "one_liner": "We find flaws in RL environments and evals and help you fix them.", "team_size": 3, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1786732497, "tags": [ "Developer Tools", "Machine Learning", "Reinforcement Learning", "B2B", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/tokenless", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/tokenless.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/usetokenless/", "x_url": "https://x.com/usetokenless", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/7f250ec6cc81104148d8bd5bafdd6b96bc77d023.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/579511aa3dc1cfb92c5f713907f31dd6fe36c640.png" ], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 3697324, "full_name": "Rohit Agarwal", "title": "Founder", "founder_bio": "Founder at Tokenless (YC S26). CS PhD student at Princeton PLI, Hazan Lab, Google Deepmind. Previously at Duolingo and theory research at Berkeley, was EECS #1 BS grad in 2024.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rohit-agarwal-b647951ba/", "x_url": "https://x.com/Rohit_Writes", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/86b74b1209c81d1d158af5587de149a3e8c2292c.jpg" }, { "user_id": 3706728, "full_name": "Andrew Liu", "title": "Founder", "founder_bio": "Founder at Tokenless (YC S26). Formerly a researcher at UC Berkeley working on language documentation tools. BA in Computer Science, Mathematics, and Linguistics from UC Berkeley. Previously TA'd CS61C.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/andrew-liu-b50873261/", "x_url": "https://x.com/landrew7363", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/04017660afe1efc138b242cbcc7c4efffc5f5bc4.jpg" }, { "user_id": 3732665, "full_name": "Kevin Wang", "title": "Founder", "founder_bio": "Founder at Tokenless. Prev MTS at SkyPilot building multi-cloud GPU infra for top neolabs, Founding Engineer at Eventual from seed to Series A building the data engine for physical AI, EECS at UC Berkeley.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/kevinzhiwang/", "x_url": "https://x.com/kevzwang", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/1a21a2cf6802cfdcf6769408ee9d25a254858622.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111407, "title": "Touchy 🙌: Voice-first AI Assistant for iPhone that understands the world around you", "tagline": "Your personal AI assistant that sees what you see, remembers what matters to you, and gets things done across all your tools. Just ask out loud.", "body": "**TL;DR:** We’re building Touchy 🙌, a voice and image-first AI assistant for the iPhone that personalizes itself to you. You can activate Touchy from your lock screen to instantly get things done across your apps or get answers about the world around you. Touchy is free to try, get it here: \n\n\n\n## **The Problem**\n\nYou probably know the feeling:\n\nYou ask the assistant on your phone, across the room, to send a colleague a message.\n\nIt cheerfully replies that it found 10 search results for “text Bob on Slack”.\n\nSo you trudge to your phone, find the app, open the DM, and begrudgingly type out your message instead.\n\nWe wanted a better personal assistant, one that had the intelligence and integrations we’ve come to expect from our AI agents, with the convenience of one that was native to your devices. We also believe that for AI to truly be personal, it needs to be deeply integrated into your life and the world around you. Right now, AI systems are constrained largely to using digital tools, yet the physical world is rich with audio and visual information that your phone is capable of receiving.\n\n## What We Built\n\nSo we built Touchy. One button press and you're instantly talking to an assistant that can see what you see, remember what you've said, and take action across all the tools you use. \n\nWhat sets Touchy apart:\n\n* **Reimagined UX.** Touchy is a fast, fully iOS-native app with a unique interface that is delightful to use and brings it closer to the world around you.\n* **Sees and hears the world around you**. Speak to Touchy and it can speak back. Easily snap a photo with the app’s camera to give Touchy access to what you’re seeing.\n* **Gets things done across your apps**. Touchy comes with built-in integrations with 1,000+ popular apps, as well as iOS-native capabilities such as you contacts and location.\n* **Make it yours over time**. You chat with Touchy in a single, continuous conversation. It automatically collects memories and preferences, and tailors itself to be more useful to you as you use it.\n\nIt's built to give you your attention back. Ask it something, let it act, and get back to your day — no app fighting for your attention.\n\n![uploaded image](/media/?type=post&id=111407&key=user_uploads/3697324/23ce7916-fce5-44c4-9ebf-f04049fee639)\n\n## The team\n\nWe’re Rohit, Andrew, and Kev. Before this, Rohit was doing a PhD on AI alignment at Google DeepMind Princeton and graduated #1 in Berkeley EECS. Andrew built educational software that scaled to half a million students worldwide. Kev was a startup founding engineer, developing data and GPU infrastructure that power many of the leading AI neolabs today. \n\nWe met during undergrad at UC Berkeley, and we’re coming back together to research, productionize, and scale the future of personal technology.\n\n## Our ask\n\nIf you have an iPhone, download Touchy and try it out: \n\nWe would love to learn what people find themselves using Touchy for, and what they wish they could use it for. You can also report bugs by shaking your phone or just telling Touchy. You can also reach us at [founders@touchyapp.com](mailto:founders@touchyapp.com)", "url": "https://www.ycombinator.com/launches/Syt-touchy-voice-first-ai-assistant-for-iphone-that-understands-the-world-around-you", "votes": 24, "created_at": "2026-08-20T21:00:00.530Z" } ] }, { "id": 33052, "name": "Mireye", "slug": "mireye", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c96db3bdd6828a4a145c678555e120d2ab43c855.png", "website": "https://www.mireye.com/", "all_locations": "San Francisco, CA, USA", "long_description": "AI agents are starting to act in the real world, but they can't reliably understand it. \r\n\r\nEvery fact about a place has its own silo: land in GIS maps, power in utility portals, people in census tables, policy in county PDFs. All built for humans, none for agents. Underneath, the data is fragmented across thousands of sources, unstructured and complex.\r\n\r\nMireye is building the infrastructure that powers agents to make decisions and act in the real world. One API and MCP to connect agents to the physical world with accurate data enrichment, signals and tools.\r\n\r\nOur mission is to index every inch of the earth, and make it as queryable as the web.", "one_liner": "Search Engine for the Physical World", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1781399078, "tags": [ "Geographic Information System", "AI", "Location-based" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/mireye", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/mireye.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/mireye/", "x_url": "https://x.com/MireyeHQ", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c96db3bdd6828a4a145c678555e120d2ab43c855.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/8abe6ac3299d5834289c40c772d5ea77b3d0f976.png" ], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 793651, "full_name": "Ansh Chokshi", "title": "Founder/CEO", "founder_bio": "Founder and CEO at Mireye (YC S26). Previously led AI/ML engineering at Seismic. At 21, co-founded Poker Pit, scaled 0 to $250K ARR in 4 months before acquisition. University of Toronto - CS, Math and Econ.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ansh-chokshi/", "x_url": "https://x.com/ansh_chokshi", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f6bfb9eddae5876264339f22a448da4cf57c8784.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108365, "title": "Mireye: The easiest way to build agentic applications for the physical world", "tagline": "One API for data, enrichment, tools and signals.", "body": "**TL;DR: Mireye is building the infrastructure that powers agents to make decisions and act in the real world. Today’s agents are book smart, we are going to make them street smart.**\n\n\n\n**The Problem:**\n\nEvery fact about a place has its own silo: land in GIS maps, power in utility portals, people in census tables, policy in county PDFs.\n\nAll built for humans, none for agents. Underneath, the data is fragmented across thousands of sources, unstructured and complex.\n\nEven with clean data the problem isn't solved, because a decision takes more than retrieving facts.\n\n**What We Built: One API and MCP that connects your agent to the physical world**\n\n![uploaded image](/media/?type=post&id=108365&key=user_uploads/793651/dcca6a7c-aafb-4c9d-89dd-8de73c0a0a04)\n\n**How It Works:**\n\nInstall our skill or paste this in your agent: \n\nAsk any question about any US location, get a citable answer in seconds\n\nData enrichment, tools and signals that turn facts into decisions, so agents can run real workflows: screen a site, underwrite a property, qualify a lead etc.\n\n300+ fields across the land, the location, the people, the rules, the risks, growing 20% week over week.\n\n**Our first customers** built agents that source off market land deals for **data centers 100x** faster than before.\n\n**You can do the same for your industry and we'll help you build it.**\n\n**Today we are launching the first ever on-demand indexing API for the physical world.** Ask for data that isn't indexed, and our long running agent sources it, and indexes it on the fly.\n\n[**Ansh Chokshi**](https://x.com/ansh_chokshi) sold his first startup, Poker Pit, at the age of 21 after growing it from zero to $250K ARR in four months. He then led AI/ML engineering at Seismic.\n\n![uploaded image](/media/?type=post&id=108365&key=user_uploads/793651/8acef036-d553-4e19-9760-961a1f90e595)\n\n**Our Ask:**\n\nAnyone building agents for physical-world applications. Comment and we'll set you up with our MCP.\n\nIntros to developers: real estate, data center, industrial, solar, nuclear.\n\nAnyone who needs data about the real world and can't get it. We'll source it and index it for you.\n\nEmail: [ansh@mireye.com](mailto:ansh@mireye.com)", "url": "https://www.ycombinator.com/launches/SBp-mireye-the-easiest-way-to-build-agentic-applications-for-the-physical-world", "votes": 340, "created_at": "2026-07-31T12:14:03.241Z" } ] }, { "id": 33053, "name": "Shiraz AI", "slug": "shiraz-ai", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/93d2ffb0a884b8a3781f59266189d89f4b30d975.png", "website": "https://shiraz.ai", "all_locations": "San Francisco, CA, USA", "long_description": "We deploy robots that learn new tasks from a single human demonstration in the factory. We're starting in contract manufacturing, where lines change every week and both traditional automation and today's VLAs fall over.\r\n\r\nOur founding team has 20+ years of experience in robotics. Navid was part of the core team that productized Spot at Boston Dynamics and led manipulation projects there. Before that, he was at Rethink Robotics. Mojtaba led whole-body control and manipulation efforts at Amazon Robotics and was a Research Assistant Professor at Northwestern University.", "one_liner": "Robots that learn on the job", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1786517990, "tags": [ "Generative AI", "Robotics", "Manufacturing" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/shiraz-ai", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/shiraz-ai.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/shiraz-ai/", "x_url": "https://x.com/shiraz_ai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/93d2ffb0a884b8a3781f59266189d89f4b30d975.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 585458, "full_name": "Navid Aghasadeghi", "title": "Founder/CEO", "founder_bio": "Founder at Shiraz AI (YC S26). Previously Senior Staff Robotics Engineer at Boston Dynamics, where he was part of the core team that brought Spot from R&D to commercial deployment and led industrial manipulation. Previously Principal Robotics Engineer at Rethink Robotics, where he developed full-stack controls for Sawyer and led force control development. Also founded GenZArt, an AI consumer app that scaled to 1M users. PhD in ECE from UIUC.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/navid-aghasadeghi-356b077b", "x_url": "https://x.com/realnavida", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b08976791c20954393282468d399741e7e6a659f.jpg" }, { "user_id": 2722145, "full_name": "Mojtaba Mozaffar", "title": "Founder/CTO", "founder_bio": "Founder at Shiraz AI. Previously Senior Applied Scientist and tech lead at Amazon Robotics, leading high-impact work in perception, policy learning, and manipulation. Before that, Research Assistant Professor at Northwestern directing AI and robotics research. PhD in Mechanical Engineering and Robotics from Northwestern.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/mojtaba-mozaffar/", "x_url": "https://x.com/MozaffarMojtaba", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2eac2883bfdf8e59573461acba516e0ea179ed74.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110367, "title": "Shiraz AI: Robots that learn on the job", "tagline": "Deploy robots that learn new tasks from a single human demonstration in the factory", "body": "**TL;DR:** We deploy robots that learn new tasks from a single human demonstration in the factory. A new task shows up on the floor? You just film it once on your phone and the robot picks it up. No sitting around collecting data. No engineer flying out to your site. No overnight fine-tuning. You show it, and the robot learns.\n\n\n\n**❌ The problem**\n\nManufacturing is full of constantly changing, high-mix production tasks, which are among the hardest to automate. A line that's packing one product today is building something completely different next week. Every factory owner we talk to says the same two things: they can't find enough people to do the work, and the automation solutions they've tried have been too expensive and too rigid, often ending up gathering dust in a corner.\n\n**✅ What we do**\n\nWe're developing a robot foundation model whose core capability is adapting instantly to new variations of a task. You give the model a human demonstration video as context, and it bridges the embodiment gap to execute the task itself. This in-context learning is the missing piece in today's VLA approaches, and it's what changing manufacturing environments need.\n\nWe deliver the complete system: robot hardware, foundation model, and on-site deployment. We're already deploying robots at customer sites.\n\n**👋 Who we are**\n\n![uploaded image](/media/?type=post&id=110367&key=user_uploads/585458/0832cd2d-2dfc-41b0-8de6-d29897b6c8d3)\n\nBetween the two of us, we have 20+ years of experience in robotics. Navid was part of the core team that productized Spot at Boston Dynamics and led manipulation projects there; before that, he was at Rethink Robotics. Mojtaba led whole-body control and bimanual manipulation efforts at Amazon Robotics and was a Research Assistant Professor at Northwestern University.\n\n**🙏 Our ask**\n\nDo you have a manufacturing task that needs automating, or know someone who does? Let us know and we'll automate it.\n\nReach us at: [founders@shiraz.ai](mailto:founders@shiraz.ai)", "url": "https://www.ycombinator.com/launches/Si7-shiraz-ai-robots-that-learn-on-the-job", "votes": 63, "created_at": "2026-08-12T20:00:00.186Z" } ] }, { "id": 33054, "name": "Lamb Labs", "slug": "lamb-labs", "former_names": [ "LAMb Intelligence" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c5ba47505a760393e85b91acdb4b3933cf8df0a7.png", "website": "http://lamb-labs.com", "all_locations": "San Francisco, CA, USA", "long_description": "Lamb Labs is building Model Processing Units (MPUs), custom AI chips that hardcode LLM weights directly into silicon. Rather than adapting general-purpose processors to run AI models, we're rethinking how LLMs are implemented in hardware to make inference dramatically faster and more energy-efficient.\r\n", "one_liner": "Lightning fast chips with hardcoded AI models", "team_size": 0, "industry": "B2B", "subindustry": "B2B", "launched_at": 1785288921, "tags": [ "Hardware", "B2B", "Semiconductors", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/lamb-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/lamb-labs.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/lamb-labs/", "x_url": "https://x.com/LambLabs", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c5ba47505a760393e85b91acdb4b3933cf8df0a7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/59a091ae426afcd919acf389592f0f3391ed83c8.png" ], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 726152, "full_name": "Niki Kotecha", "title": "Founder/CEO", "founder_bio": "Founder at Lamb Labs. AI PhD at Imperial College London + I-X Scholar. MEng/BA in Engineering from University of Cambridge. Alan Turing Institute Global Fellowship and 160+ citations", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/niki-kotecha/", "x_url": "https://x.com/KotechaNiki", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9d7169b597815c0e4b02ff12606c162b4c12432b.jpg" }, { "user_id": 3179989, "full_name": "Thomas Lanning", "title": "Founder/CTO", "founder_bio": "Founder at Lamb Labs. Masters in Mathematical and Theoretical Physics from Oxford. Top of his class in undergraduate at Edinburgh University. (Re)discovered the Higgs-boson using ML and CERN data. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/thomas-lanning/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/aadf202207708e0feff18ade84592a81efb19f2f.jpg" } ], "jobs": [ { "id": 116541, "title": "Founding AI Accelerator Architect (Research & Hardware Co-Design)", "url": "https://www.ycombinator.com/companies/lamb-labs/jobs/QfZbNOf-founding-ai-accelerator-architect-research-hardware-co-design", "location": "London, England, GB", "type": "Full-time", "role": "Engineering", "role_type": "Hardware", "salary_range": "£60 - £160 GBP", "equity_range": "", "experience": "Any (new grads ok)", "visa": "US citizenship/visa not required", "skills": [] }, { "id": 116543, "title": "Founding ASIC Engineer (Architecture & RTL)", "url": "https://www.ycombinator.com/companies/lamb-labs/jobs/cW2pB1O-founding-asic-engineer-architecture-rtl", "location": "London, England, GB", "type": "Full-time", "role": "Engineering", "role_type": "Electrical", "salary_range": "£60 - £160 GBP", "equity_range": "", "experience": "1+ years", "visa": "US citizenship/visa not required", "skills": [] } ], "news": [], "launches": [ { "id": 108778, "title": "Lamb Labs: Custom Chips for AI Inference", "tagline": "Co-designed models and silicon: we hardcode the model architecture into chips targeting 20,000+ tokens per second and 63x higher intelligence per watt.", "body": "**Hi Everyone!** \n\n**We’re Niki and Thomas - cofounders of Lamb Labs.**  \n\n**TL;DR: We build custom AI inference chips that run inference at 20,000+ tok/s and 63x higher intelligence per watt than a traditional GPU. Fully local and private by construction, no subscriptions, no API costs.** \n\n**Our mission: the world’s fastest LLM at the lowest power.** \n\n\n\n**Our hot take: AI models are finally getting good enough to hardcode them directly into silicon.** Once you stop treating the chip as a general-purpose GPU and burn a specific model into hardware, inference gets radically faster and cheaper per watt. Most of the industry is still building flexible, general chips. We think that flexibility is exactly what's wasting your energy. So we’re on a mission to reduce the energy footprint of AI. \n\n**The Problem**\n\nAI is taking power on two fronts: concentrating control of data and intelligence, and consuming ever more electricity to grow. That future scares us, so we're building against it.\n\nHere's the technical root of the waste: **GPUs burn most of their energy during inference moving data, not computing.** LLM/VLM inference is memory-bound, fetching a weight from memory costs \\~100–1000x the energy of the arithmetic you do with it. So you underutilize the GPU you paid for: some GPUs only hit 20–40% compute utilization during LLM inference while the rest idles waiting on memory.\n\nCloud hosting adds its own tax: expensive per-token bills, latency you can't control, network dependencies, and your data leaving your control.\n\n**Our Solution**\n\nWe build the model and the chip together. \n\nThe Hardware Layer\n\n* **Custom architecture, RL-tuned.** We built an RL environment that co-designs the chip architecture for a given model, optimizing directly for speed and energy per token.\n* **Aggressive quantization.** Removes complex math, which shrinks the size and cost of the chip.\n* **Weights on-chip.** The model weights live where the compute is overcoming the memory-bandwidth bottleneck\n\n \n\nBecause everything is hardcoded, it's fully local and secure by construction. No connectivity dependency, nothing leaves the device.\n\nThe Software Layer \n\nDesigning the chip led us to solving a fundamental flaw in LLMs so we are now open-sourcing a new model next week, converted from another open source model, to run 2x faster on any hardware. Come back next week for part 2 of our launch to find out more….\n\n**Our Ask**\n\nIf you answer \"yes\" to any of these, we'd love your help:\n\n* **Are you paying for inference in watts, not just dollars?** Neocloud or private DC operators who are power- or hardware-cost-bound rather than demand-bound, talk to us.\n* **Do you need inference where the cloud isn't an option?** Private AI deployments, regulated, air-gapped, sovereign, or bandwidth-starved deployments.\n* **Do you want to deploy the fastest AI models in your stack?** High frequency trading, agentic workflows \n* **Are you stuck on a thermal or battery budget?** Robotics, wearables, consumer AI hardware where a Jetson-class part is over-provisioned and runs too hot.\n* **Can you introduce us to anyone above?** Warm intros are the single most useful thing.\n\nContact us at [contact@lamb-labs.com](mailto:contact@lamb-labs.com)", "url": "https://www.ycombinator.com/launches/SIU-lamb-labs-custom-chips-for-ai-inference", "votes": 64, "created_at": "2026-08-03T04:56:33.889Z" } ] }, { "id": 33055, "name": "Capto", "slug": "capto", "former_names": [ "chromie.dev", "Chromie" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/300579f704f4971abc355f226a91f0da7f7f38b4.png", "website": "https://gocapto.com/", "all_locations": "San Francisco, CA, USA", "long_description": "The intelligent operating system for government contracting. Built by AI engineers and former government officials.", "one_liner": "Build something the government wants", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1785887406, "tags": [ "GovTech", "Workflow Automation" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/capto", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/capto.json", "year_founded": null, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/chromie-dev/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": "https://github.com/chromie-dev", "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/300579f704f4971abc355f226a91f0da7f7f38b4.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/e7b5908337f1787dba0fe86b95ef9cbb88b6f3db.png" ], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 2373564, "full_name": "Ananth Sankaralingam", "title": "Founder/CEO", "founder_bio": "Founder at GoCapto.com, bridging the gap between Silicon Valley and the US Government.\n\nPreviously studied AI/ML University of Maryland, published ML research at NeurIPS, and worked as an engineer at LinkedIn and Amazon. \n\nOutside of work, I love sports and music. I've completed a half iron man, perform Carnatic Music (I play the Mridangam, a South Indian drum), and post on my YouTube channel.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ananth-s/", "x_url": "https://x.com/_ananthhh", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/808def4baf8fc8cb01cf72462f93ef1e5c32654c.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109881, "title": "Chromie - AI OS for Government Contracting", "tagline": "Find, qualify, & win government contracts with AI", "body": "**TL;DR:** Chromie is the AI operating system for government contractors. We help teams find opportunities before the RFP drops, build the right relationships, test win strategies against real data, and draft winning proposals. All on top of a company brain that remembers everything your team has ever learned.\n\n\n\n\\\n**The problem**\n\nThe U.S. government is the largest buyer on earth, spending $800B on contracts every year. Yet the business is fundamentally relationship driven. Winning depends on understanding the agency, finding opportunities earlier than your competitors, and applying judgment built over years. \n\nToday, that work runs on gut instinct, fragmented tools, and institutional knowledge that disappears when people leave.\n\n**What Chromie does**\n\nChromie surfaces agency signals and expiring contracts years before an RFP is posted publicly, giving capture teams time to understand the customer and shape a credible pursuit.\n\n![uploaded image](/media/?type=post&id=109881&key=user_uploads/2373564/7796de7e-c1c5-4a18-b1ce-7036de7414dc)\n\nFrom there, teams can qualify opportunities, develop teaming and pricing strategies, and compare potential win strategies without wasting resources. Chromie’s PWin models are backtested against millions of verified award outcomes, helping teams augment intuition with evidence.\n\nOnce a team decides to bid, Chromie carries that strategy into proposal development. It maintains a **company brain** that connects capture, proposal, and contract-execution knowledge in one shared foundation, so each pursuit benefits from the company’s past experience.\n\n![uploaded image](/media/?type=post&id=109881&key=user_uploads/2373564/899638cb-afd8-40d6-81c5-e625eba27b75)\n\nWhat your team learns today strengthens the next pursuit.\n\n**Traction**\n\nChromie already supports contractors with 100+ prime awards and more than $250M in obligated contract value. These teams have used Chromie to 3x their monthly proposal output and win new awards.\n\n![uploaded image](/media/?type=post&id=109881&key=user_uploads/2373564/8292fe04-c346-49ba-9a3c-836b09c75f60)\n\n**Who we are**\n\nWe're Ananth and Akshay. We met in a high school English class and have been building together ever since. Between us we've built and scaled infrastructure at Google, LinkedIn, Plaid, and Amazon. We’re developing Chromie alongside former federal contracting officers and acquisition leaders with decades of government experience.\n\n**Our ask**\n\nWe'd love to connect with:\n\n* Government contractors, especially capture, BD, and proposal teams\n* Startups and companies trying to break into government contracting \n* GovCon and federal-sales experts willing to share what they’ve learned :)\n\nContracts are won by the companies that start earliest, remember most, and test how to win instead of guessing. If that's you, or you're trying to get there, reach out.\n\nLearn more at [chromie.dev](http://chromie.dev), or email [founders@chromie.dev](mailto:founders@chromie.dev) to find your company's fit in government contracting 🧩\n\n![uploaded image](/media/?type=post&id=109881&key=user_uploads/2373564/26ff0c87-9129-4af4-b345-f797f269e874)\n\n", "url": "https://www.ycombinator.com/launches/SaH-chromie-ai-os-for-government-contracting", "votes": 24, "created_at": "2026-08-10T17:48:31.293Z" } ] }, { "id": 33056, "name": "83 Sciences", "slug": "83-sciences", "former_names": [ "LabAlly", "83 sciences" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/08c94ab9737bf76c36b1cbac477c55b5067788ed.png", "website": "https://83sciences.ai", "all_locations": "San Francisco, CA, USA", "long_description": "83 Sciences uses AI to mine discarded experimental data and drive scientific breakthroughs. Working alongside labs, we discover the materials that will power the new Industrial Revolution.", "one_liner": "AI-native materials discovery powered by unpublished experimental data", "team_size": 3, "industry": "Industrials", "subindustry": "Industrials", "launched_at": 1783362788, "tags": [ "Hard Tech", "Advanced Materials", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/83-sciences", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/83-sciences.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/83sciences.ai", "x_url": "https://x.com/83sciences", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/08c94ab9737bf76c36b1cbac477c55b5067788ed.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/0671a743b093a1e19c485330f4a4b3434448a3af.png" ], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 2982957, "full_name": "Ian Naccarella", "title": "Founder/CEO", "founder_bio": "Founder at 83 Sciences developing the future of scientific discovery. Previously a fellow at MIT focused on commercializing IP out of the ecosystem, a strategy manager at Sila Nanotechnologies (battery materials unicorn), and a consultant at Boston Consulting Group. Hold a BS and MS in Chemical Engineering from Stanford and MBA from Harvard Business School.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ian-naccarella", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a9328006522cbb0cbf9cea515c5a0324457d8b7f.jpg" }, { "user_id": 3679894, "full_name": "Eric Riesel", "title": "Founder/CTO", "founder_bio": "Founder at 83 Sciences. Experimental chemist and ML researcher with a PhD in inorganic chemistry from MIT, where I built machine learning systems to analyze complex materials data in minutes instead of days. My research spans quantum materials, generative models, and experimental data analysis, including one of the first generative approaches for chemical experimental data. BA in Chemistry from Columbia.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/eric-riesel", "x_url": "https://x.com/ericriesel", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6bc3086b3dce349c94c3909537c479266d39939c.jpg" }, { "user_id": 3380797, "full_name": "Yankang Yang", "title": "Founder/COO", "founder_bio": "Founder at 83 Sciences. Former BCG Principal; led the firm's AI program scaling tools and AI practices to 30,000+ users globally. Harvard CS & EE undergrad. Past robotics research (Wyss Institute) and biosignals research (Roswell Park) experience ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/yankangyang/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d0a82615e5de87cad9ef0ba0f984d1d3b0dac538.jpg" } ], "jobs": [ { "id": 98793, "title": "Founding AI Engineer", "url": "https://www.ycombinator.com/companies/83-sciences/jobs/Wgkya40-founding-ai-engineer", "location": "San Francisco, CA, US / New York, NY, US / Remote (US)", "type": "Full-time", "role": "Engineering", "role_type": "Machine learning", "salary_range": "$120K - $250K", "equity_range": "0.50% - 5.00%", "experience": "1+ years", "visa": "US citizen/visa only", "skills": [] }, { "id": 102160, "title": "Head of Data Partnerships", "url": "https://www.ycombinator.com/companies/83-sciences/jobs/NR8uCdC-head-of-data-partnerships", "location": "New York, NY, US / San Francisco, CA, US / Remote (US)", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "$170K - $220K", "equity_range": "0.50% - 1.00%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 108805, "title": "83 Sciences: Materials discovery from the ~90% of experimental data that never gets published", "tagline": "We train materials discovery models on experimental data to help industry improve manufacturing performance at scale and help academics publish more. ", "body": "History’s greatest ages are named after materials (Stone Age, Bronze Age, Iron Age, Silicon Age, etc.). The next great age is ours to make together.\\\n\\\n\n\\\n**TL;DR:** We are a team of Stanford, MIT, and Harvard researchers discovering breakthrough materials by training frontier models on decades of unpublished experimental data. **We've already discovered a novel material and drafted a manuscript for publication.**\n\nIndustrial R&D teams use us to find scalable application-specific materials that frontier models can't predict. Academic labs turn decades of unpublished data into new papers, citations, and commercialization revenue with us.\n\n**What we do**\n\nWe are uncovering novel materials that help industry improve performance at scale and stay top of new regulations. Examples include replacements for PFAS, the \"forever chemicals\" that are being banned worldwide, and finding a binder for silicon anodes that can survive charging and drop into existing battery manufacturing processes (our cofounder Ian’s work at Sila Nanotechnologies).\n\n**Why we're different**\n\nThe materials performance that matters to industry can only be measured in real experiments, but >85% of experimental research is never published. Frontier models train on papers and simulations and miss all the failures sitting in file drawers. We train on that unpublished data through partnerships with labs across the country, and it’s already working: we've already discovered a novel material and drafted a manuscript for publication.\n\n**Who we are**\n\nWe are Eric, Ian, and Yankang, collaborators for years on everything from superconductors to cooling textiles, and we lead a team of world-class researchers and commercialization experts.\n\nEric developed one of the early generative AI approaches for analyzing experimental data (MIT PhD), Ian has brought multiple climate technologies to market, and Yankang led Boston Consulting Group’s AI program, scaling frontier technology to 30,000+ users globally. Advisors: Simon Billinge (UCSB), Mingda Li (MIT), Ashwin Sreenivas (Decagon).\n\n**Our ask**\n\nIf you need better materials to drive performance and cost, especially in batteries, semis, membranes, catalysts or pharma, or know an academic lab sitting on experimental data, we'd love an intro: [hello@83sciences.ai](mailto:hello@83sciences.ai)", "url": "https://www.ycombinator.com/launches/SIv-83-sciences-materials-discovery-from-the-90-of-experimental-data-that-never-gets-published", "votes": 16, "created_at": "2026-08-03T22:41:26.834Z" } ] }, { "id": 33059, "name": "Palisade", "slug": "palisade-2", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/15205df0821b37ecf5cd82136bd682c2cb65c1b2.png", "website": "https://palisade-ai.com/", "all_locations": "San Francisco, CA, USA", "long_description": "", "one_liner": "AI that finds and fixes OS-level vulnerabilities across device fleets", "team_size": 1, "industry": "B2B", "subindustry": "B2B -> Security", "launched_at": 1784742432, "tags": [ "SaaS", "B2B", "Cybersecurity", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Security" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/palisade-2", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/palisade-2.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/palisade-ai/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/15205df0821b37ecf5cd82136bd682c2cb65c1b2.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 2688178, "full_name": "Fnu Prince", "title": "Founder", "founder_bio": "Left Caltech at 18 to start Palisade (YC S26). Previously did red-teaming at 16, USACO Platinum/USAPhO HM/AIME, built my own OS for fun (pre-Claude/Codex)", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/fnu-prince-a98405276/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/382f016fead85bdbb69a24e2165f092473c60b9c.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 106995, "title": "Palisade - AI that detects and fixes OS vulnerabilities across every endpoint in minutes", "tagline": "AI that finds and fixes OS-level vulnerabilities across device fleets", "body": "**TL;DR:** For companies running large fleets of OS-level endpoints (running GPU clusters, having a network of cloud devices, or deploying on-device), Palisade detects and fixes operating-system vulnerabilities (privilege escalation, remote code execution, memory corruption) that are extremely hard to detect and mitigate without applying AI to kernel-level data (syscalls, eBPF traces, process/memory events, network packets). A fix that used to take weeks to roll out, Palisade now applies in minutes across every endpoint.\n\nHey YC! I'm Prince, an ML researcher and engineer who dropped out from Caltech at 18 to build Palisade.\n\n**What’s the problem:** Modern attacks increasingly target the operating system, but most security products only observe the network, cloud, or application layers. That leaves OS vulnerabilities among the hardest problems to detect and the slowest to remediate across a fleet.\n\n**Why I’m building this:** Since I was 16, I have done plenty of professional red-team and blue-team security work. One pattern kept showing up: OS-level vulnerabilities were consistently the hardest to handle. Detection required low-level visibility, and even after identifying a vulnerability, rolling out a mitigation across hundreds or thousands of machines often took weeks. My background in both machine learning and operating systems convinced me this didn't have to be a manual process, so I built Palisade to automatically detect and remediate these vulnerabilities at the kernel level across an entire fleet.\n\n**What Palisade does:** Palisade watches your systems at the kernel level—syscalls, eBPF traces, process and memory events, and network packets—and uses AI to find OS vulnerabilities other tools miss, then implement the fix automatically across every endpoint.\n\n![uploaded image](/media/?type=post&id=106995&key=user_uploads/2688178/fa8c6d6a-b058-4f44-87d0-5511d038718b)\n\n_The Palisade control plane, live against a real kernel (LSM-BPF). The live feed at the bottom: every row is a real syscall decision._\n\nThe vulnerabilities it catches are the ones you can't see from the app or network layer alone:\n\n* Privilege escalation\n* Remote code execution\n* Memory corruption\n\nMost are caught before an exploit ever happens, and the fix is applied immediately, fleet-wide. The same patch that used to take weeks now lands in minutes.\n\n![uploaded image](/media/?type=post&id=106995&key=user_uploads/2688178/2f4e756d-1348-44a2-bdce-bb3491925da5)\n\n_The host audit: real findings ranked by severity. The bottom one is the cloud-metadata exposure: the same path behind the Capital One breach. One click on **Apply fix** and it's installed as a kernel rule._\n\nPalisade can already detect and fix real problems:\n\n* CVE-2024-1086: use-after-free in the Linux kernel's netfilter (nf_tables); local privilege escalation to root\n* CVE-2025-6019: libblockdev/udisks privilege escalation to full root\n* CVE-2025-32463: sudo --chroot local privilege escalation to root\n\n**Who it's for:** Companies running a lot of OS-level endpoints, where a single kernel vulnerability multiplies across the whole fleet. Whether it’s network servers or GPUs, we have you covered.\n\n\n\n**Our asks:**\n\n* Intros: if you run a large fleet of OS-level endpoints, or know a security/infra lead who does, I'd love an intro.\n* See the demo and tell me if you’d use it. Brutal feedback welcome.\n\n— [prince@palisade-ai.com](mailto:prince@palisade-ai.com), , DMs welcome", "url": "https://www.ycombinator.com/launches/Rpj-palisade-ai-that-detects-and-fixes-os-vulnerabilities-across-every-endpoint-in-minutes", "votes": 8, "created_at": "2026-07-23T00:33:01.523Z" } ] }, { "id": 33062, "name": "Hubble", "slug": "hubble-ai", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/8d58ef930283a4163ee9544b3936abd1327a463a.png", "website": "https://hubble.ai/", "all_locations": "Seattle, WA, USA", "long_description": "Hubble’s mission is to unify the patient medical record and remove the barriers to accessing it. Today, a person’s health history is scattered across disconnected systems, leaving patients and the people helping them to piece it together. We’re building a future where one authorization unlocks a more complete record, regardless of where someone received care.\r\n\r\nOur founders previously built healthcare and AI products at Grow Therapy and Amazon One Medical, and we’re backed by Y Combinator, Madrona, and Basis Set. We’re building our team in Seattle, with the opportunity to work directly with the founders and shape how people access and use their health information.", "one_liner": "AI-native retrieval engine for medical records", "team_size": 2, "industry": "Healthcare", "subindustry": "Healthcare -> Healthcare IT", "launched_at": 1784648214, "tags": [ "B2B", "Healthcare" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare", "Healthcare IT" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/hubble-ai", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/hubble-ai.json", "year_founded": 2026, "city": "Seattle", "country": "US", "linkedin_url": "https://www.linkedin.com/company/hubble-ai-health/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/8d58ef930283a4163ee9544b3936abd1327a463a.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 3661311, "full_name": "Prabha Dublish", "title": "Founder/CEO", "founder_bio": "Founder at Hubble. Previously first PM at Grow Therapy launching their first AI features & PM at Meta.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/prabhadublish/?skipRedirect=true", "x_url": "https://x.com/prabhadublish", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8065deceef33a65a9ee1389e3ac45816e2e0d754.jpg" }, { "user_id": 3662596, "full_name": "Aaron Leon", "title": "Founder/CTO", "founder_bio": "CTO at Hubble. Previously tech lead at Amazon One Medical, where he launched Amazon's first healthcare AI agents serving 1M+ patients. Prior, led data platforms across Amazon Health.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/aaronleon1/", "x_url": "https://x.com/aaronleon_x", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/065b8c9d2baf322d16f034c345fbb7ff591f5e50.jpg" } ], "jobs": [ { "id": 114400, "title": "Founding Engineer ", "url": "https://www.ycombinator.com/companies/hubble-ai/jobs/0ATelWV-founding-engineer", "location": "Seattle, WA, US", "type": "Full-time", "role": "Engineering", "role_type": "Backend", "salary_range": "$150K - $175K", "equity_range": "1.00% - 3.00%", "experience": "3+ years", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 109532, "title": "Hubble: Retrieve medical records other APIs can't", "tagline": "Your user verifies their identity once, and we return their medical records through one API.", "body": "**TL;DR:** Hubble gives your product a patient's medical records through a single API. Your user verifies their identity once, and we assemble their records from across the providers they've seen and return them for your AI agent to use. Where APIs are not sufficient, we keep going with AI (browser and voice agents). Every request runs on the patient's individual right of access.\n\n**The problem**\\\nIf you're building something that needs a patient's records, there's no clean way to get them. A patient's history is scattered across every provider they've ever seen, and each system is its own island. Most APIs returns the patient-portal subset rather than the full record, and coverage is inconsistent because every health system sets up its integration differently. So most teams either build per-provider integrations for months, fall back to faxes and manual portal logins, or push the problem back onto the patient by asking them to track down their own records or remember a login they set up years ago and have long since lost.\n\n**What Hubble does**\\\nYour user verifies their identity once, and we go get their records so you don't have to.\n\n* **One verification, no work on your end.** The patient verifies their identity in a couple clicks. After that, assembling the records is our job, not yours and not theirs.\n* **We don't stop at the portal subset.** We connect into EHRs and HIEs where they work, and when an API comes back empty we keep going.\n* **Browser and voice agents for everything else.** Agents submit the patient's request through provider portals and call records departments on the patient's behalf, working the phone trees so nobody on your team has to.\n* **Built on the patient's right of access.** Every request runs on the patient's individual right of access. They review every source we request from and can revoke at any time. Every retrieval is logged.\n\n**About the team**\\\nWe're the team behind the AI agents at Grow Therapy and Amazon's One Medical, used by millions of patients. Prabha was Grow's first PM building and scaling their AI agents. Aaron built the AI infrastructure at One Medical serving 1M+ clients. (Fun fact: we've known each other for over a decade.)\n\n**Our ask**\n\n* If your product needs patient records, reach out and we'll get you access: [prabha@hubble.ai](mailto:prabha@hubble.ai)\n* Intros to anyone building AI products on patient data, in digital health, care delivery, life sciences, or legal tech.\n\nSign up at [hubble.ai](http://hubble.ai)\n\n", "url": "https://www.ycombinator.com/launches/SUe-hubble-retrieve-medical-records-other-apis-can-t", "votes": 21, "created_at": "2026-08-10T16:00:00.087Z" } ] }, { "id": 33111, "name": "Skymerse", "slug": "skymerse", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/dcac201e21fa70f9b3a8a1b827c8fd7759c25920.png", "website": "https://skymerse.com", "all_locations": "San Francisco, CA, USA", "long_description": "Skymerse is building one AI system for airline flight operations: from planning and monitoring in the operations center to real-time support in the cockpit. It is developing specialized aeronautical models that understand the data, constraints, and procedures behind each flight, replacing fragmented tools and manual interpretation. Notamify, its first product, already serves airline customers and thousands pilots and aviation professionals.", "one_liner": "Autopilot for Flight Operations", "team_size": null, "industry": "Industrials", "subindustry": "Industrials -> Aviation and Space", "launched_at": 1783640155, "tags": [ "Airlines", "Aerospace", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Aviation and Space" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/skymerse", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/skymerse.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/skymerse", "x_url": "http://x.com/skymerse", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/dcac201e21fa70f9b3a8a1b827c8fd7759c25920.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 1472269, "full_name": "Damian Szumski", "title": "Founder/CEO", "founder_bio": "Founder at Skymerse (YC S26). Licensed Flight Dispatcher. Previously Product Owner at Flightradar24 API. Background in Aerospace and Flight Ops.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/damian-szumski/", "x_url": "https://x.com/zaxrev", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/720059d9d72b5293515142d82e45f90f7f3e7eb7.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111421, "title": "Skymerse: Autopilot for Flight Operations", "tagline": "From planning and monitoring in the operations center to real-time support in the cockpit.", "body": "I’m Damian, founder of Skymerse (S26). I’m a licensed flight dispatcher and have spent 10 years working across flight operations, aviation data, and software.\n\nWe’re building one AI system to automate flight operations - from planning and monitoring in the operations center to real-time support in the cockpit.\n\nOur goal is full automation of the work behind every flight.\n\n\n\n## The problem\n\nAirlines have spent decades digitizing flight operations. The information moved from paper to screens. The work of understanding it did not. A flight exists across many systems. Each system may be good at its own job, but the complete operational picture exists between them.\n\nPilots and operations teams still have to build that picture themselves. They search, compare, and interpret information to plan, monitor, and adapt every flight from schedule to landing.\n\nWhen something changes, they do it again. When responsibility moves from the operations center to the crew, much of the context has to be passed on or rebuilt.\n\nToday, people are the integration layer.\n\n## What Skymerse does\n\nSkymerse builds a live operational model of every flight. When an airline’s schedule flows in, it assembles the operational context for each flight and keeps it current as conditions change.\n\nThat model is the foundation for automation. Skymerse does not stop at displaying information or sending an alert. It understands what a change means for the flight, updates the operation, and carries the required work forward automatically.\n\nSkymerse plans, monitors, and adapts flights from the operations center to the cockpit. Instead of rebuilding context across tools and handoffs, Skymerse manages each flight as one continuous process, from schedule to landing.\n\n## Why I started Skymerse\n\nSkymerse started with [Notamify.com](http://Notamify.com). I built Notamify because I was amazed by how much information pilots and ops teams have to process before every flight - and how much of that work was still done by hand.\n\n![uploaded image](/media/?type=post&id=111421&key=user_uploads/1472269/7e9fe835-e22f-4cb1-92dc-ac0f99a0cce2)\n\nNotamify interprets every single NOTAM into operational impact. It reduces the briefing time from 2 hours to 2 minutes per flight. It now serves airline customers and thousands of pilots and aviation professionals.\n\nAs we built it, the larger problem became clear. Making one source easier interpreted is valuable, but people still have to connect it with the rest of the operation. They still have to work out what it means for their flight and what to do next.\n\nSkymerse is that next step: from helping people understand one part of a flight to automating flight operations e2e.\n\n## Automate your flight operations\n\nWe are onboarding airlines that want to automate their flight operations. If you lead flight operations, dispatch, or an OCC, we want to show you Skymerse on your flights.\n\n[Book a demo](http://skymerse.com/demo) or email [hello@skymerse.com](mailto:hello@skymerse.com).", "url": "https://www.ycombinator.com/launches/Sz7-skymerse-autopilot-for-flight-operations", "votes": 30, "created_at": "2026-08-20T23:00:00.215Z" } ] }, { "id": 33144, "name": "Grip", "slug": "grip", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/a6be2671c5d4e3adfe916a4a59aad4680c7cc9d3.png", "website": "https://www.griprobotics.ai", "all_locations": "Zürich, ZH, Switzerland", "long_description": "Grip is a robotics company building intelligent robotic systems for waste and recycling. We are starting with organic waste, where removing contaminants from messy, constantly changing streams remains difficult and highly manual. Our technology combines learning-based manipulation and custom hardware to identify and pick objects, learning from real-world operating data, including every successful and failed grasp, to become more reliable and adaptable over time. This helps waste operators reduce manual sorting, recover more valuable material, lower contamination, and deploy automation flexibly across facilities. Over time, Grip aims to expand beyond organics and build the intelligent manipulation layer for trash in general, making waste processing and the circular economy more efficient and scalable.", "one_liner": "Robots for waste manipulation", "team_size": 4, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1785254669, "tags": [ "Robotic Process Automation", "Sustainability" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "Switzerland", "Europe" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/grip", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/grip.json", "year_founded": 2026, "city": "Zürich", "country": "CH", "linkedin_url": "https://www.linkedin.com/company/griprobotics", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/a6be2671c5d4e3adfe916a4a59aad4680c7cc9d3.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/f0ba95cdc010ecb5489de28455277fd61e5501df.png" ], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3678212, "full_name": "Jonas Gruetter", "title": "Founder/CEO", "founder_bio": "Former robotics researcher, published at ICRA. Ex-Gravis Robotics, Robotic Systems Lab, and Ursa Mining. Now building robots to solve fundamental societal problems.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jonasgrutter/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/81ff3bd631bc4b7206041d530bc182993fddd8fc.jpg" }, { "user_id": 3678142, "full_name": "Stanislaw Piasecki", "title": "Founder/CTO", "founder_bio": "Co-founder at Grip. Previously founding engineer at URSA Mining, building autonomous excavators with custom control circuits, RL simulators, and multi-sensor perception systems. Built autonomous interceptor drones at CDDS, autonomous race cars at AMZ and quadruped robotics at ANYbotics. MS in Computational Science & Engineering from ETH Zürich.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/stanislaw-piasecki/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/253c9593fed9e59498847d4c158543099816d8f1.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111317, "title": "Grip - Robotics for Waste Management", "tagline": "Robots that sort any waste and recover recyclables, learning from every grasp.", "body": "Hi everyone - Stan and Jonas here, founders of **Grip**. We build intelligent robots that sort _any_ waste, not just the clean, predictable objects existing systems are limited to. We're starting with plastic contamination in organic waste.\\\n\\\n\n\n**The problem**\n\nWe didn't start with a thesis, we found it on sorting lines, after visiting waste facilities across several countries:\n\n* **Contamination is existential.** Plastic that organic-waste facilities can't remove gets shredded, processed, and spread onto agricultural soil. Operators spend up to 30% of operating costs fighting it and still fall short. Even Germany's 0.1% limit for plastic in finished compost is a tacit admission that full removal isn't possible with today's tools.\n* **The labor force is aging and shrinking.** The picking work is unpleasant and turnover is brutal. Operators increasingly aren't choosing between people and robots. They're choosing between robots and nobody.\n* **Dangerous items cause fires.** Batteries, propane tanks, and vapes slip through and ignite during handling and shredding: 448 reported fires at US and Canadian waste facilities in 2025, \\~$2.5B in damage, and insurers are walking away. At the same time trash sorting is sitting in the top 5 of US deadliest occupations.\n\n**What we built**\n\nThe hard part isn't the gripper. It's that waste never looks the way you planned. Objects arrive crushed, folded, tangled, half-buried, and moving. Put today's sorting robots in front of this and they break.\n\nWe're building one manipulation system that picks whatever shows up, at the front of the line where the material is most chaotic and existing automation stops. That means chaotic scene understanding, robust grasp planning, and hardware built for real waste streams: our custom three-way gripper is designed to maximize throughput, taking out as much contamination per cycle as possible.\n\nOur rugged hardware matters, but intelligence is the product. Every pick generates data no other system ever encounters: what the robot saw, what it tried, how the material responded, whether it worked.\n\n![uploaded image](/media/?type=post&id=111317&key=user_uploads/3678142/bc867fdc-9225-4b85-83f9-f83433465d5b)\n\n**Why now**\n\nWaste policy has moved from vague ambition to enforceable deadlines:\n\n* Separate bio-waste collection is now mandatory across the EU. More material, and more contamination, every year.\n* New EU fertiliser rules just cut permitted impurities in compost and digestate, with national limits like Germany's stricter still.\n* The US is real: California mandated a 75% cut in landfilled organics, missed it, and enforcement is ramping. At the same time, landfills near major cities are filling up, and that pressure is pushing more and more US municipalities to collect organic waste separately. More material, more contamination, more facilities that need to sort it.\n\nMore than 6,500 biowaste plants and thousands of recycling facilities across Europe and North America face this with no robot that can touch it. That's Grip.\n\n**Why us**\n\nWe've built autonomous race cars, quadrupeds, drones, and excavators, robots that operate in dirt, weather, and chaos rather than controlled labs. Grip is that experience applied to an industry where physical intelligence removes a real bottleneck.\n\n**Our ask**\n\nWe'd love intros to:\n\n* Owners and operators of composting, anaerobic-digestion, and organic-waste facilities\n* Owners and operators of MRFs and recycling facilities\n* Regulators and experts shaping contamination limits and recycling standards\n\nIf someone runs a facility and will show us their hardest sorting problem, we'll get on a plane.\n\n[**griprobotics.ai**](http://griprobotics.ai)", "url": "https://www.ycombinator.com/launches/SxR-grip-robotics-for-waste-management", "votes": 15, "created_at": "2026-08-19T13:00:00.941Z" } ] }, { "id": 33177, "name": "HyperProbe", "slug": "hyperprobe", "former_names": [ "HyperTest inc" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/b4b91e0b78bbed1c83add57d22b06ab632567450.png", "website": "https://hyperprobe.co", "all_locations": "San Francisco, CA, USA", "long_description": "With tools like Cursor and Claude Code, writing software is no longer the bottleneck. Debugging is. Agents cannot fix production crashes using logs and traces that were sized for human eyeballs. The gap is missing data.\r\n\r\nHyperProbe allows agents to debug live software autonomously. Our SDK drops non-breaking probes straight into running services to read exact variable states from memory at the moment of failure, instead of reasoning backwards from flat logs.\r\n\r\nZero redeploys. Under 1% CPU overhead. Automatic PII redaction.\r\n\r\nWe turn 4-hour production war rooms into 10-minute automated resolutions. As agents write and operate more of the world's software, HyperProbe is the runtime observability layer that lets them understand what they built.\r\n\r\n", "one_liner": "AI engineer that debugs and fixes production incidents in realtime", "team_size": 8, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1783455456, "tags": [ "AIOps", "Developer Tools", "SaaS", "DevOps", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/hyperprobe", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/hyperprobe.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/b4b91e0b78bbed1c83add57d22b06ab632567450.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3173209, "full_name": "Shailendra Singh", "title": "Founder/CEO", "founder_bio": "Founder at HyperProbe. Also founded HyperTest. Previously headed and scaled engineering and product functions at a unicorn and few small startups.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shailendra-singh-6540b8b/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/804a582f290d73d89dadb5a73815fc0814ca121d.jpg" }, { "user_id": 417680, "full_name": "Karan Raina", "title": "Founder/CTO", "founder_bio": "Founder and CTO at HyperTest (YC S26)\nMS CS Georgia Tech\n\n2nd time founder\nPreviously led a 50 member engineering team at Limetray", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/karanraina/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/bd379360253464d14cb5b5a47a3911c2139eaa6d.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110401, "title": "Hyperprobe: On-call Agent for engineering teams", "tagline": "The fastest way for coding agents to debug incidents live in production.", "body": "# \\*\\*\\*\\*\n\n**TL;DR:** AI writes most of our code now. Debugging it when it breaks in production is still slow, because the value that explains the failure was never logged. HyperProbe is an on-call agent for engineering teams. It takes an open incident to a confirmed root cause by dropping read-only probes into your running service and capturing the exact variable state logs don't have. No redeploys, no restarting the app. \\\n\\\n📹 **Launch Video: [https://youtu.be/QtaZrQ8_Uk8](https://youtu.be/QtaZrQ8_Uk8)**\n\n**🚧 The Problem**\n\nTeam ships faster than ever with agents that write most of our code. This is already shrinking the useful context engineers need to debug AI written code. Not helped by the limited telemetry added in the same code by AI. So when something breaks in prod, the first instinct for engineers is to open logs or throw them to agents. But if the line you are looking for is not there, agents will start guessing the root cause on limited data, forcing engineers to collect more logs by redeploys.\n\nThe result: our best engineers become our on-call team, incidents stay open longer, and patches come from data that never capture production context, so issues recur.\n\nYou can see where that leads. In September last year, a public fintech in India paid out $5 million after failed transactions were incorrectly marked successful. NYSE the same year miscalculated price bands for 40 symbols, printing trades so wrong they had to be reversed. Same pattern: the code ran clean, wrote the wrong state, and neither logs nor traces could show it.\n\n ✅ **What Hyperprobe does**\n\nHyperProbe gives engineering teams an on-call agent that works the incident and takes it to a confirmed root cause, live in production.\n\n* **Plan:** the agent picks up the alert from PagerDuty, Datadog, or Slack, then reads your logs and traces to locate the file and line. The easy incidents resolve right here.\n* **Probe:** logs not enough? places read-only virtual breakpoints on the suspect lines and captures the exact variable state live, across services. No redeploys, no restarts. \n* **Confirm:** the diagnosis is verified against real captured evidence, not a guess. Your engineer gets a confirmed RCA and commits the fix.\n\nThere are two pieces. An SDK that runs inside your service, and an MCP server your coding agent talks to. The SDK is what makes setting probes (virtual breakpoints, log or metric) possible without a redeploy. \n\n### **🔍 A few examples**\n\n* **An intermittent 500 that won't reproduce locally:** catch it live with full variable state \n* **A third-party API misbehaving:** see the exact payload and response\n* **A function returns a wrong value with no error:** see what it computed and why\n* **Anything that only happens in prod:** stop redeploying with console.logs\n\n🚀 **The team**\n\nWe've both lived this problem. Shailendra ran production at scale. Karan was a senior architect at a SaaS company. Before this, both spent three years building production SDKs for a testing product, learning the hard parts of pulling real runtime state out of a running service without breaking it.\n\nEvery incident we watched came down to the same question: what was the actual value at the moment it broke. Nobody built it for the way engineers debug with coding agents now. So we did.\n\n![uploaded image](/media/?type=post&id=110401&key=user_uploads/3173209/ddf3d8d0-2bd4-4a8c-9547-fa5ea2496469)\n\n### **🎁 Why now**\n\nFor thirty years the expensive part of software was writing it, and every tool got built for that. Generation is now nearly free, but the cost didn't disappear. It moved downstream, to figuring out why the generated code broke in production.\n\nAnd that system is now mostly code no human wrote or holds in their head. When it breaks, the engineer has less context than ever, and their first move is already to hand it to the coding agent. That agent only has your logs and traces, so it guesses. We give it the evidence instead, captured live from the running code at the point of failure.\n\n### **🙏 Our ask**\n\n* If you're an engineer tired of the redeploy loop, debugging code you didn't write at 2 AM, try it on your own service.**,** first investigation in 10 min:[ https://docs.hyperprobe.co/quickstart](https://docs.hyperprobe.co/quickstart)\n* If your best engineers have become your on-call team and you want that fixed, let us work one real incident from your stack, free: ", "url": "https://www.ycombinator.com/launches/Sif-hyperprobe-on-call-agent-for-engineering-teams", "votes": 82, "created_at": "2026-08-17T15:30:20.732Z" } ] }, { "id": 33178, "name": "TovenAI", "slug": "tovenai", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/2b832a1e965bc59dfbc18a9817802870d7264332.png", "website": "https://toven.ai", "all_locations": "San Francisco, CA, USA", "long_description": "TovenAI builds a platform of AI agents that handle manual, data-intensive work for compliance teams at institutional trading firms, with six agents currently live. Our customers range from specialist firms to large global institutions, with our agents supporting multiple compliance workflows at a top-10 institutional brokerage.\r\n\r\nWith prebuilt connectors to more than 30 systems—including Bloomberg, Cloud9, and NICE Actimize—firms can onboard to our platform and evaluate results using their own data in less than a week.", "one_liner": "AI agents for compliance at institutional trading firms", "team_size": 2, "industry": "Fintech", "subindustry": "Fintech -> Banking and Exchange", "launched_at": 1786560317, "tags": [ "Fintech", "Compliance", "Regtech", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech", "Banking and Exchange" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/tovenai", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/tovenai.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/tovenai/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/2b832a1e965bc59dfbc18a9817802870d7264332.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 695373, "full_name": "Batu Balci", "title": "Founder/CEO", "founder_bio": "Founder at TovenAI. Previously Quantitative Research Engineer at Citadel's Global Quantitative Strategies group. Senior Software Engineer at Salesforce. Stanford CS.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/batuhan-balci-7a3589114/", "x_url": "https://x.com/batubbalci", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/24f0a192733f83d6c638efaa5efeb8c9521f2e84.jpg" }, { "user_id": 2817439, "full_name": "Sam Turchetta", "title": "Founder/CTO", "founder_bio": "Cofounder at TovenAI. Previously worked on ML at Meta. Stanford Men’s Varsity Tennis and CS ’21.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/sam-turchetta-b14311166/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/25fb2bee3c76413e5285ab143882be546aabb6c8.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110467, "title": "TovenAI: Your intelligent compliance stack for finance", "tagline": "AI agents for compliance at institutional trading firms that understand firm policies, analyze activity across systems, and surface what matters most", "body": "**TL;DR:** Compliance teams at institutional trading firms still spend hours manually sifting through mountains of data across recorded calls, chats, and trade records to investigate activity and produce defensible audit trails. Determining what matters requires not only finding the right data, but also interpreting activity against applicable regulations and the firm’s own policies and procedures. TovenAI is building a platform of AI agents that connects to their systems, assembles the evidence, and surfaces what matters for compliance teams at these firms—think the sales and trading businesses at banks like Bank of America and specialist brokerages like Marex.\n\nWe are live at one of the top 10 institutional trading firms in the U.S., with multiple AI agents delivering value through our platform.\n\n\n\n## **What we built**\n\nTovenAI’s agents analyze activity across existing communications and trading systems using each firm’s policies, procedures, and applicable regulations as context, assembling the relevant evidence for compliance teams.\n\nOur trade-reconstruction agent rebuilds the full sequence of a trade, links every event to its source evidence, and reduces the process from around four hours to minutes. Our electronic-communications surveillance agent reduces the false-positive burden by 3× and, in a head-to-head evaluation, delivered roughly 7× more value per alert reviewed than the market-leading incumbent.\n\nOur platform supports several other compliance workflows, giving teams broader coverage with less manual work.\n\nWith prebuilt integrations to more than 30 systems – including Bloomberg, Cloud9, CME and NICE Actimize – firms can evaluate TovenAI on their own data in under a week without replacing their existing stack.\n\n## **Why us**\n\nWe met eight years ago at Stanford. Batu later worked as a quantitative research engineer at Citadel Global Quantitative Strategies, building AI systems for quantitative research. Sam was a machine learning engineer at Meta, working on its ads recommendation systems.\n\nWe work closely with our customers and sit besides compliance officers every day. This gives us a rapid feedback loop, lets us see their workflows firsthand, and allows us to redesign compliance operations with AI at their core.\n\nSam’s mother spent her 15-year career in risk and compliance and was our first sounding board as we began exploring the space.\n\n## **Our ask**\n\nWe would love introductions to compliance leaders at firms with institutional trading businesses.\n\nThink the sales and trading divisions of banks like **JPMorgan, Bank of America, Citizens**, as well as specialist brokerages like **TP ICAP and Marex**. We are interested in firms of all sizes, from specialist brokerages to the largest global banks.\n\nIf you know a broker or trader at one of these firms, they can likely connect us with someone on their compliance team.\n\nA one-line introduction is enough, and we will take it from there.\n\nEmail us at: [**founders@toven.ai**](mailto:founders@toven.ai)\n\nLearn more about us at .", "url": "https://www.ycombinator.com/launches/Sjj-tovenai-your-intelligent-compliance-stack-for-finance", "votes": 63, "created_at": "2026-08-19T18:58:36.456Z" } ] }, { "id": 33179, "name": "Verdict Machine", "slug": "verdict-machine", "former_names": [ "TxAI" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/efc3c74e529cc296ebd75ff470655d005300ddf0.png", "website": "https://verdictmachine.com", "all_locations": "Tel Aviv-Yafo, Tel Aviv District, Israel", "long_description": "Verdict Machine uses AI to understand each financial institution's complete on-chain environment, assess its cyber risk and generate tailored controls. The same live model powers transaction prevention, continuous cyber diligence, security posture and cost-efficient wallet screening. Before execution, Verdict Machine evaluates proposed transactions against the institution's active policies.\n\nIn our first weeks, we started POCs with asset managers and digital-asset infrastructure providers whose systems move billions of dollars every week for financial institutions. We're now in active evaluations with digital-asset teams at some of the world's largest banks.\n\nThe founding team launched and led blockchain security at Check Point (NASDAQ: CHKP).", "one_liner": "AI cybersecurity for financial institutions' digital assets", "team_size": 4, "industry": "B2B", "subindustry": "B2B -> Security", "launched_at": 1786374737, "tags": [ "Crypto / Web3", "Security", "AI" ], "tags_highlighted": [ "Crypto / Web3" ], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Security" ], "regions": [ "Israel", "Middle East and North Africa" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/verdict-machine", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/verdict-machine.json", "year_founded": 2026, "city": "Tel Aviv-Yafo", "country": "IL", "linkedin_url": "https://www.linkedin.com/company/verdict-machine/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/efc3c74e529cc296ebd75ff470655d005300ddf0.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 85721, "full_name": "Dan Danay", "title": "Founder/CEO", "founder_bio": "Co-founder and CEO at Verdict Machine (YC S26). Previously Head of Blockchain Security at Check Point (NASDAQ:CHKP), launching their digital assets division. Founded keyTango (MPC-powered crypto clearing house) and Airmada (YC W16, automated drone stations). MIT Sloan MBA. Captain in Israeli Navy, commanded missile ship and naval electronics school.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dandanay/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/821ae7bdee6621edf11f400f8aee456ae9e5b649.jpg" }, { "user_id": 3784711, "full_name": "Isaac Grossman", "title": "Founder", "founder_bio": "Co-founder and CTO at Verdict Machine (YC S26). Previously VP Engineering at Fhenix, where he led development of an FHE-enabled L2, and at SCRT Labs, where he helped scale Secret Network to a top-100 crypto asset. Co-founded keyTango, an MPC-powered crypto clearing house.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/itzik-grossman", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/729b4a4e5f449d069ff7dde45b6a9d7af5cb6e7c.jpg" }, { "user_id": 3784793, "full_name": "Liran Kogan", "title": "Founder/CPO", "founder_bio": "Co-founder and CPO at Verdict Machine (YC S26). Previously launched and led blockchain-security products at Check Point (NASDAQ: CHKP), and led payments and FX products at Payoneer (NASDAQ: PAYO). NYU + INSEAD.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/liran-kogan/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b8ffca19f9687957470301ec575fb6f9cab05b65.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110493, "title": "Verdict Machine: AI cybersecurity for financial institutions’ digital assets", "tagline": "AI that understands each institution's complete on-chain environment, assesses its cyber risk, and generates tailored controls to prevent hacks", "body": "Hey all! We are Dan, Liran, and Isaac. Before Verdict Machine, we launched blockchain security at Check Point (NASDAQ: CHKP).\n\n**TL;DR:** Blockchain is becoming critical infra for all finance. More than $3.4B in digital assets was stolen in 2025. Verdict Machine is AI cybersecurity for financial institutions' digital assets. We use AI to understand each institution's environment, assess cyber risk and generate tailored controls that prevent hacks.\n\n**Building for:** banks, asset managers, DeFi protocols and infra, companies that use stablecoins or RWAs. If that’s you, email us at [founders@verdictmachine.com](mailto:founders@verdictmachine.com).\n\n\n\n### Signed is not safe\n\nSophisticated hackers don’t need to compromise a wallet to steal funds. They only need to trick an authorized operator into approving one malicious transaction.\n\nOn-chain, settlement is final. There is no chargeback and no reversal. The security decision has to happen before the transaction executes.\n\nAttackers are sophisticated and see the entire institution, while existing security tools see fragments:\n\n* Hidden dependencies. Vendors, protocols, bridges and infrastructure are assessed separately or not at all.\n* Limited context. Controls check a transaction, not the institution behind it.\n* Stale controls. Rules are written once and reviews run once, while the institution and its risk keep changing.\n\n \n\n### How Verdict Machine works\n\nVerdict Machine uses AI to secure each client end to end.\n\n**Inventory.** It builds the complete inventory of a client’s on-chain environment: assets, wallets, smart contracts, users, vendors, counterparties, and transaction patterns.\n\n**Posture.** It quantifies cyber risk and recommends tailored controls to minimize exposure.\n\n**Enforce controls.** Before execution, it evaluates each proposed transaction to prevent attacks in real time.\n\nThe same live model powers every workflow:\n\n**Real time prevention.** Security and compliance teams secure their most critical transactions.\n\n**Continuous cyber diligence.** Asset managers and risk teams underwrite the cyber risk of protocols, vendors, and counterparties before and after capital is deployed.\n\n**Wallet screening.** Payments and stablecoin teams screen for fraud, unauthorized transactions, and risky wallets.\n\n### Why now\n\nInstitutions are racing to launch stablecoin and tokenized-asset products. MiCA and DORA are in force, so the people who own this risk now have to show their work. Attackers punish under-protected organizations.\n\nFireblocks and Chainalysis reached $8B valuations for digital asset custody and compliance. Cybersecurity for on-chain finance is a bigger opportunity. Every financial institution is going to use digital assets and therefore will require digital-asset cybersecurity.\n\n### Traction\n\nWe left Check Point the day before Y Combinator started. We already have POCs with asset managers and digital-asset infrastructure providers moving billions of dollars weekly for financial institutions, and we are in active evaluations with a digital-asset team at one of the world's largest banks.\n\n### Why us\n\nWe previously launched and led blockchain security at Check Point (NASDAQ: CHKP). Our CTO Isaac is a cyber veteran from Matzov, Israel’s cryptography unit. We are crypto natives with many years building infra for crypto and fintech. We built Verdict Machine to bring AI-native, institution-specific security to on-chain finance.\n\nManaging or deploying crypto, stablecoins, or tokenized assets? Talk to us.\n\n[verdictmachine.com](http://verdictmachine.com) · [founders@verdictmachine.com](mailto:founders@verdictmachine.com)", "url": "https://www.ycombinator.com/launches/Sk9-verdict-machine-ai-cybersecurity-for-financial-institutions-digital-assets", "votes": 24, "created_at": "2026-08-17T15:00:01.008Z" } ] }, { "id": 33248, "name": "Dock", "slug": "dock", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/94f67a762d3f3ae2966b1300c5b2c0c8e7e007de.png", "website": "https://trydock.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Dock is where you hire AI agents for business. Bring an idea, and your agents turn it into a company that runs: they find the customers, do the work and keep it moving, day and night. Building a company used to mean hiring a building full of people first. With Dock, entrepreneurs can start a real business today and run it like a company of a hundred. You decide where it goes. Your agents make it happen.", "one_liner": "AI agents that run your business.", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1785286566, "tags": [ "B2B", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/dock", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/dock.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/trydockai", "x_url": "https://x.com/trydock", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/dock-74d7", "github_url": "https://github.com/try-dock-ai", "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/94f67a762d3f3ae2966b1300c5b2c0c8e7e007de.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 31632, "full_name": "Mike Molinet", "title": "Founder/CEO", "founder_bio": "Co-founder at Dock. Previously co-founder at Thena.ai and co-founder/COO at Branch.io, where we scaled to $100M+. Stanford MBA, Bucknell Engineering.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/mikemolinet/", "x_url": "https://x.com/mikemolinet", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/7b574771b9b6bda3dde6f3e9d8054e5d524816a7.jpg" }, { "user_id": 3679532, "full_name": "Govind Kavaturi", "title": "Co-Founder", "founder_bio": "Co-founder of Dock (YC S26), helping entrepreneurs start and run businesses with AI agents. Previously co-founded Thena, early at Branch.\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/govindkavaturi/", "x_url": "https://x.com/Govikavaturi", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/8fd52c5b64bfbbb6b4e333f1e7c91d26cc6b4e7f.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 114146, "title": "Dock – A multiplayer agent workspace that gives you a team of proactive agents that feel like colleagues instead of chat boxes", "tagline": "Get 3x more done in a fraction of the time with a team of Dock agents.", "body": "After a decade building and scaling companies — including Branch, which[ Govind](https://www.linkedin.com/in/govindkavaturi/) and I grew from zero to $100M+ ARR — we kept hitting the same wall: doing more always meant hiring more. Now, for the first time, it doesn't have to. That's why we built[ **Dock**](https://trydock.ai/).\n\n\n\nAI agents have changed what's possible, which is why we're building again. Dock gives you a team of proactive AI agents  that work alongside you — they talk to each other, hand off work, and take initiative instead of waiting to be prompted. So you steer instead of babysitting a dozen chat windows.\n\n## **The Problem**\n\nAI agents can finally do real work. But today they work alone — isolated chats that sit there waiting for you to prompt them. The moment you run more than one, _you_ become the middleman: copy-pasting context between a dozen windows and unblocking every agent by hand. The more agents you add, the more work you create for yourself. It doesn't scale.\n\n## **The Solution**\n\nDock turns those isolated agents into a real team. Here's how it works:\n\n### **1. Spin up a team of specialized AI agents in minutes.**\n\nEach agent has a real role, its own objectives, and persistent memory — so it gets better at its job over time instead of resetting every session.\n\n### **2. They work together, proactively.**\n\nYour agents  pursue goals on their own, message each other, and hand off work when one needs something from another — no copy-pasting, no you relaying context. They only come back to you when there's a real decision to make.\n\n### **3. You and your team work right alongside them.**\n\nEveryone — humans and AI agents  — works in the same shared workspace, toward the same goals. The earlier you plug Dock in, the more your agents  learn about how your company actually works.\n\n## **Why this works**\n\nMulti-agent coordination is an infrastructure problem, and we built that infrastructure. Dock runs its own agent harness and runtime, executes in the cloud or on your machine, and gives every agent memory, schedules, and a shared workspace as primitives. Agents working as a team is not a feature we added. It’s what the system is.\n\n## **Why we're building this**\n\nWe've spent our careers inside growing companies — hiring, scaling, managing — and felt the coordination drag of every new person you add. When AI made it possible to grow output without growing the team, this problem was already ours. In a real way, Dock found us more than we found it. Today we run our entire company on Dock, and we think soon every company will run on humans and AI agents  working together.\n\nGet early access:[ **trydock.ai**](http://trydock.ai)\n\n![uploaded image](/media/?type=post&id=114146&key=user_uploads/31632/f4079ba2-afab-462d-b130-e3e6215af412)\n\n", "url": "https://www.ycombinator.com/launches/Th4-dock-a-multiplayer-agent-workspace-that-gives-you-a-team-of-proactive-agents-that-feel-like-colleagues-instead-of-chat-boxes", "votes": 9, "created_at": "2026-09-09T16:00:00.120Z" } ] }, { "id": 33249, "name": "OpenRelay", "slug": "openrelay", "former_names": [ "VectorLay" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/9c5dcebff59e38045b8e976d964cbe43ec732a9e.png", "website": "https://www.openrelay.inc", "all_locations": "San Francisco, CA, USA", "long_description": "", "one_liner": "Distributed, hardware-agnostic AI inference", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1783565051, "tags": [ "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/openrelay", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/openrelay.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/openrelayinc/posts/", "x_url": "https://x.com/openrelayinc", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/openrelay-inc", "github_url": "https://github.com/OpenRelayInc/OpenRelay", "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/9c5dcebff59e38045b8e976d964cbe43ec732a9e.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 1876109, "full_name": "Jaden Wang", "title": "Founder/CEO", "founder_bio": "ex Voltage Park Engineering HPC, ex TensorDock, Founder@Heaviside Compute, UW Dropout", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jaden-wang-b99327177/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/82ebee97bf559c8af23a9617985f3b19baa98db7.jpg" }, { "user_id": 3742866, "full_name": "Prashant Patel", "title": "Founder/CTO", "founder_bio": "Founder at OpenRelay (YC S26). Previously Staff Engineer at Voltagepark building managed Inference and Orchestration software for 10,000+ GPU clusters and serverless compute that scaled to 10M+ executions/month. Founding member of Amazon Bedrock at AWS, led Custom Model Import from 0 to multi-million ARR and scaled inference infrastructure for models up to trillions of parameters. MS CS from NYU.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/prashant182/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/0a8e85496afb951cba03393a0de6f6a76b102057.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109526, "title": "OpenRelay: The Inference Delivery Network", "tagline": "One API for inference across every chip.", "body": "**TL;DR:** OpenRelay is building an Inference Delivery Network (IDN) on top of its CDN for GPUs. We provide one entrypoint for inference across every chip. \n\nSend us a workload and we’ll cut your costs up to 20% by running it through our network, balancing the best available accelerator (NVIDIA, TPU, Trainium, AMD) across clouds and handing back a single endpoint. You never pick the hardware, chase quotas, or rebuild your stack per provider. And if you run GPUs or other AI accelerators (an NVIDIA cluster, TPUs, a reserved fleet), you can plug your capacity into our rails and we'll bring the inference demand. Get started [here](https://openrelay.inc/).\n\nP.S. sign up for our Luma page to be notified of our Launch Party!\n\n\n\n\\\n\n\n## The Team\n\nHey guys! We're **Jaden and Prashant**, the founders of OpenRelay. We came at the same problem from opposite sides.\n\n* Jaden built a half-megawatt data center at 20 from a warehouse shell, then joined Voltage Park early, where he ran virtualization and helped stand up HPC.\n* Prashant built accelerator infra at AWS, where he ran custom model deployments and wrote custom kernels to squeeze out Trainium performance.\n\n \n\nWe kept seeing the same thing: there's more AI compute than ever, and you still can't reach it. We left our jobs to fix that.\n\n## The Problem\n\nCompute is everywhere; liquidity is nowhere. Accelerators are scattered across dozens of clouds, chip vendors, and operators, each behind its own quotas, drivers, contracts, and console. Teams that want to run inference can't reach the capacity, and the capacity can't reach them.\n\nRunning inference naively is easy; running it at scale is hard. Any operator can spin up a model on a box. Turning idle GPUs into a production endpoint (load balancing, isolation, autoscaling, failover, metering, billing) is a software platform most individual operators won't build. So capacity sits stranded behind their own front door, and developers either overpay for scarce reserved GPUs or hand-stitch a fragile multi-provider stack.\n\n## Solution\n\nWe built one set of rails across all of it.\n\nSend a workload (any container or model) and we schedule it onto the best-fit accelerator, attach a production endpoint, and scale it. We route to the cheapest accelerator that meets your latency and throughput targets, and we benchmark continuously. If the market can't beat renting yourself, we backstop with our own inference demand, so you’re never on the hook. \n\nYou never pick the chip, cloud, or region; we handle routing, isolation, failover, metering, and billing across 4+ clouds and every major chip family. We're language- and framework-agnostic (no SDK lock-in), and you can drive everything from our CLI and REST API.\n\nIt's two-sided: if you run GPUs or other accelerators (a neocloud, a data center, a pay for an underutilized reserved fleet), you can connect that capacity to our rails and we'll bring the inference demand, turning hardware into a revenue-generating endpoint without building a platform yourself.\n\n## Traction\n\nWe're **live in production**, generating 100 billion tokens a week across 22 physical locations across Europe, APAC, North America, and the Middle East running on 8 different accelerator SKUs.  \n\n## Asks\n\nWe'd love to hear from you!\n\n* **Run GPUs, or know someone who does?** This is who we most want to talk to: companies running their own nodes for inference, data centers, reserved fleets, even a crypto or mining cluster looking to move into AI. Plug your capacity into our rails and we'll bring the demand. Reach us at [founders@openrelay.inc](https://mailto:founders@openrelay.inc).\n* **Shipping inference and tired of picking chips?** Bring a workload and we'll get you on the best available hardware today.\n* **Feedback, intros, or just want to chat?** Book a call with us, or email [founders@openrelay.inc](mailto:founders@openrelay.inc) for anything else.", "url": "https://www.ycombinator.com/launches/SUY-openrelay-the-inference-delivery-network", "votes": 12, "created_at": "2026-08-07T17:00:00.601Z" } ] }, { "id": 33277, "name": "Hop Aero", "slug": "hop-aero", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/a5cab9ede70e23b0ea3f8e3b64b7a19735939c17.png", "website": "https://hopaero.com", "all_locations": "Los Angeles, CA, USA", "long_description": "Hop Aero is building rocket cargo vehicles that deliver to contested environments hundreds of miles away in minutes. Our rockets fly point-to-point into the most austere locations at hypersonic speeds. Hop Aero's goal is to provide the U.S. and allies the asymmetric advantage that deters a future major conflict.", "one_liner": "Rocket cargo delivery to contested environments", "team_size": 6, "industry": "Industrials", "subindustry": "Industrials -> Defense", "launched_at": 1781029529, "tags": [], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Defense" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/hop-aero", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/hop-aero.json", "year_founded": 2024, "city": "Los Angeles", "country": "US", "linkedin_url": "https://www.linkedin.com/company/hop-aero/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/a5cab9ede70e23b0ea3f8e3b64b7a19735939c17.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 480112, "full_name": "Matija Milenovic", "title": "Founder/CEO", "founder_bio": "Founder at Hop Aero. Previously CEO at porkchop (electric propulsion for satellites), launched on SpaceX Transporter-3, built $400M+ sales pipeline, sold IP. System Engineer at Destinus leading 20+ person team on supersonic UAV. Scaled Ukrainian loitering munition production from 10 to 100+ units/month. MS Aerospace Engineering KTH Stockholm.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/matija-milenovic/", "x_url": "https://x.com/mat_milenovic", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5de9c335ad613ed7722aedbdf0a67bed625083f9.jpg" }, { "user_id": 2233348, "full_name": "Jacob Balaj", "title": "Founder/CTO", "founder_bio": "Founder at Hop Aero. Previously Integration & Test Engineer at SpinLaunch (satellites), Lead Systems Integration & Test at Masten Space Systems (lunar lander for NASA's CLPS program), and Integration & Test at Virgin Orbit (LauncherOne) and SpaceX (Falcon 9/Heavy—built booster refurbishment process, cut test time in half). Earlier avionics QA on MV-22 tiltrotors in the US Marine Corps.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jacob-balaj-6b4b061b5/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ef902d83986d7f7b76472502d93c8fe3621a2b2a.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112220, "title": "Hop Aero 🚀 Rocket Cargo", "tagline": "Hop Aero delivers an autonomous 21st century D-Day", "body": "[https://drive.google.com/file/d/1xqm08Bpwv1LXOsas-oVfZ5rDwp1lNqP8](https://drive.google.com/file/d/1xqm08Bpwv1LXOsas-oVfZ5rDwp1lNqP8/view?usp=share_link)\n\\\n❌ **Problem**\n\nWhile Jacob was serving in the U.S. Marine Corps, Typhoon Haiyan had just swept through the Philippines. Jacob and his unit were given a clear order: deliver life-saving supplies of medicine and food as soon as possible. Despite working around the clock, the effort was regularly blocked due to aircraft breaking down, flight-critical parts being out of inventory, and pilot fatigue.\n\nSimilarly, when building his previous startup, Matija’s team were given 24 hours’ notice to deliver a satellite propulsion system to the mission integrator. FedEx and UPS could not guarantee on-time delivery, forcing Matija to make a 23-hour trip to hand-deliver the flight hardware. He spent the flight back thinking that there has to be a better way.\n\nIt turned out that it wasn’t just us. Everyone from startups to the U.S. military, needs faster delivery. We saw an opportunity to create the fastest form of transportation in the world.\n\n🚀 **Rockets are the answer**\n\nWe are building Rook, an autonomous rocket that delivers 250 kg of cargo up to 750 km away. That’s Okinawa to Taiwan in 15 minutes.\n\nRook launches from a standard 40-ft shipping container, flies through space, re-enters at hypersonic speeds and lands on unprepared surfaces, allowing it to launch and land anywhere on the planet. This opens up completely new commercial markets and military capabilities.\n\nRook can deliver swarms of autonomous unmanned systems to the battlefield, delivering a fully-autonomous 21st century D-Day effect for the U.S. and allies. The same Rook can deliver time-critical goods for manufacturing, automotive, aerospace, offshore oil and pharma, preventing downtime costs of up to $2M/hour.\n\n🏆 **Traction**\n\n* **Won a $1.25M DoW contract** from the US Air Force.\n* **Completed our first tethered test** at the Oklahoma spaceport on a small-scale prototype\n* **Successfully hot-fired our engine and igniter** built using hopOS, our AI-augmented secret sauce.\n* **Successfully demonstrated drone deployment** from our rocket cargo vehicle.\n\n🥷 **Team**\n\n**Matija (CEO):** Previously founded porkchop (electric propulsion for small satellites) - launched on SpaceX Transporter-3, built a $400M+ sales pipeline, sold the IP. Led a 20+ person supersonic UAV team at Destinus and scaled Ukrainian loitering-munition production from 10 to 100+ units/month.\n\n**Jacob (CTO):** US Marine Corps (MV-22 avionics) → SpaceX (Falcon 9/Heavy; built the OG booster refurb program) → Virgin Orbit (LauncherOne) → Masten (led systems integration & test on NASA's CLPS lunar lander) → SpinLaunch. A decade of making rockets fly on schedule.\n\n![uploaded image](/media/?type=post&id=112220&key=user_uploads/480112/8946bdbe-09b5-447f-8022-6233e2401800)\n\n🙏 **Asks**\n\n* **Intros to DoW customers and end-users.** Anyone with warm paths into the Navy, PACOM, or program offices working contested logistics - we want to talk.\n* **Intros to commercial end-users.** Businesses with high downtime costs such as manufacturing, automotive, aerospace, offshore oil, pharma - all are right down our alley.\n* **Payload partners.** We want to deploy your drones, interceptors, electronic warfare jammers and more! Reach out if you want to give your product the ability to teleport to the battlefield!", "url": "https://www.ycombinator.com/launches/TC0-hop-aero-rocket-cargo", "votes": 6, "created_at": "2026-08-26T15:00:00.587Z" }, { "id": 107889, "title": "Hop Aero 🚀 Rocket Cargo", "tagline": "Rockets that deliver cargo anywhere on Earth in minutes, including battlefields", "body": "**TLDR: We’re building rockets that deliver cargo faster than anything else on earth. We already have a $1.25M contract with the US Air Force for delivery to contest environments, and we’ve completed a tethered flight of our small-scale prototype.** 🚀🌎\\\n\\\n[https://drive.google.com/file/d/1dI3bRpiuqdShHQ2xHMhqOhPluYQMGS2B](https://drive.google.com/file/d/1dI3bRpiuqdShHQ2xHMhqOhPluYQMGS2B/view?usp=sharing)\n\n❌ **Problem**\n\nWhile Jacob was serving in the U.S. Marine Corps, Typhoon Haiyan had just swept through the Philippines. Jacob and his unit were given a clear order: deliver life-saving supplies of medicine and food as soon as possible. Despite working around the clock, the effort was regularly blocked due to aircraft breaking down, flight-critical parts being out of inventory, and pilot fatigue.\n\nSimilarly, when building his previous startup, Matija’s team were given 24 hours’ notice to deliver a satellite propulsion system to the mission integrator. FedEx and UPS could not guarantee on-time delivery, forcing Matija to make a 23-hour trip to hand-deliver the flight hardware. He spent the flight back thinking that there has to be a better way.\n\nIt turned out that it wasn’t just us. Everyone from startups to the U.S. military, needs faster delivery. We saw an opportunity to create the fastest form of transportation in the world.\n\n🚀 **Rockets are the answer**\n\nWe are building Rook, an autonomous rocket that delivers 250 kg of cargo up to 750 km away. That’s Okinawa to Taiwan in 15 minutes.\n\nRook launches from a standard 40-ft shipping container, flies through space, re-enters at hypersonic speeds and lands on unprepared surfaces, allowing it to launch and land anywhere on the planet. This opens up completely new commercial markets and military capabilities.\n\nRook can deliver swarms of autonomous unmanned systems to the battlefield, delivering a fully-autonomous 21st century D-Day effect for the U.S. and allies. The same Rook can deliver time-critical goods for manufacturing, automotive, aerospace, offshore oil and pharma, preventing downtime costs of up to $2M/hour.\n\n🏆 **Traction**\n\n* **Won a $1.25M DoW contract** from the US Air Force.\n* **Completed our first tethered test** at the Oklahoma spaceport on a small-scale prototype\n* **Successfully hot-fired our engine and igniter** built using hopOS, our AI-augmented secret sauce.\n\n🥷 **Team**\n\n**Matija (CEO):** Previously founded porkchop (electric propulsion for small satellites) - launched on SpaceX Transporter-3, built a $400M+ sales pipeline, sold the IP. Led a 20+ person supersonic UAV team at Destinus and scaled Ukrainian loitering-munition production from 10 to 100+ units/month.\n\n**Jacob (CTO):** US Marine Corps (MV-22 avionics) → SpaceX (Falcon 9/Heavy; built the OG booster refurb program) → Virgin Orbit (LauncherOne) → Masten (led systems integration & test on NASA's CLPS lunar lander) → SpinLaunch. A decade of making rockets fly on schedule.\n\n![uploaded image](/media/?type=post&id=107889&key=user_uploads/480112/7e957579-e672-47be-9433-9742f3c618bc)\n\n🙏 **Asks**\n\n* **Intros to DoW customers and end-users.** Anyone with warm paths into the Navy, PACOM, or program offices working contested logistics - we want to talk.\n* **Intros to commercial end-users.** Businesses with high downtime costs such as manufacturing, automotive, aerospace, offshore oil, pharma - all are right down our alley.\n* **Payload partners.** We want to deploy your drones, interceptors, electronic warfare jammers and more! Reach out if you want to give your product the ability to teleport to the battlefield!", "url": "https://www.ycombinator.com/launches/S49-hop-aero-rocket-cargo", "votes": 11, "created_at": "2026-07-28T14:33:36.526Z" } ] }, { "id": 33278, "name": "PRINCEPS", "slug": "princeps", "former_names": [ "Princeps Dev" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/8face6fd127a729a79aa67ef1d9e4669f91060ab.png", "website": "https://princeps.dev", "all_locations": "London, England, United Kingdom", "long_description": "PRINCEPS is the first purpose-built insurance company for the trillion-dollar compute buildout. We underwrite SLA, GPU residual value, and outage risk, and price the coverage that unlocks cheaper financing for data centers and neoclouds.", "one_liner": "AI-native insurance company for the compute economy", "team_size": 2, "industry": "Fintech", "subindustry": "Fintech -> Insurance", "launched_at": 1786600478, "tags": [ "Fintech", "Machine Learning", "Insurance", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech", "Insurance" ], "regions": [ "United Kingdom", "Europe" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/princeps", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/princeps.json", "year_founded": 2026, "city": "London", "country": "GB", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/8face6fd127a729a79aa67ef1d9e4669f91060ab.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 3685149, "full_name": "Conrad Frøyland Moe", "title": "Founder/CEO", "founder_bio": "Co-Founder of PRINCEPS. Previously Treasurer of the Oxford Union (£3M+ institution). Experience from national Norwegian politics. Co-Founded an events company at whilst at Oxford University: scaled to 650+ tickets per event at £15K+ revenue. \nB.A Oxford (History & Politics)", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/conrad-frøyland-moe-7950b1229/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/3f9d3c85109c1e611b1a72db32bde5c3f770bd8f.jpg" }, { "user_id": 3671622, "full_name": "Anya Trofimova", "title": "Founder/COO", "founder_bio": "Co-Founder of PRINCEPS. Previously worked on data centre project finance, including a $2bn Gulf data centre deal and the first ever space data centre project, and built software for UK Parliament on energy and industrial policy. Youth Fellow at Chatham House researching defence and infrastructure security. Oxford Casberd academic Scholar and former VP of the Oxford Union ($3M+ institution). Designed and led a UN digital campaign at 16.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/anyatrofimova", "x_url": "https://x.com/anya_trofimova7", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/46f402a445d84bb9a5f2b87c6c18d4cc67c14ea4.jpg" } ], "jobs": [ { "id": 114920, "title": "Founding Engineer", "url": "https://www.ycombinator.com/companies/princeps/jobs/nLXB8CE-founding-engineer", "location": "London, England, GB", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$80K - $150K", "equity_range": "0.50% - 1.50%", "experience": "1+ years", "visa": "US citizenship/visa not required", "skills": [] } ], "news": [], "launches": [] }, { "id": 33279, "name": "Hebbian Robotics", "slug": "hebbian-robotics", "former_names": [ "Vanda Labs" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/4e7272f21860060f8aa33bb297af6d18fab5ffdc.png", "website": "https://hebbianrobotics.com/", "all_locations": "San Francisco, CA, USA", "long_description": "Hebbian Robotics builds open source foundational infrastructure for Physical AI to develop robots that extend what humanity can do, for the benefit of the communities they serve.", "one_liner": "Foundational infrastructure for Physical AI", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1786123276, "tags": [ "Robotics", "Data Science", "Data Engineering", "AI", "Databases" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/hebbian-robotics", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/hebbian-robotics.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/hebbian-robotics/", "x_url": "https://x.com/hbr_pbc", "fb_url": null, "cb_url": null, "github_url": "https://github.com/Hebbian-Robotics", "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/4e7272f21860060f8aa33bb297af6d18fab5ffdc.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/18513ded63751a708b6d42e2a84d2517cb404aaf.png" ], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 388120, "full_name": "Brandon Ong", "title": "Co-Founder", "founder_bio": "Brandon is the co-founder of Hebbian Robotics, which builds APIs that help robotics data teams verify data quality for model training. He previously built Gigit AI, a Sequoia-backed AI customer experience startup. His research on AI for low-resource languages topped SEA-HELM, an LLM benchmark co-developed with Percy Liang's research group at Stanford. He briefly pursued a PhD in Robotics at Columbia and NTU before starting Hebbian.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ongjjbrandon", "x_url": "https://x.com/brandon1", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5b65343ad1d0b482401872b3124da731618a3058.jpg" }, { "user_id": 968709, "full_name": "Kingston Kuan", "title": "Co-founder", "founder_bio": "Kingston spent the last few years at Jane Street building systems that scaled to millions of messages per second. He graduated valedictorian from the National University of Singapore (NUS) in Computer Engineering.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/kingston-kuan/", "x_url": "https://x.com/kstonekuan", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b674a8b80df5cd21ef03f13e6470d6126bc4c333.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109711, "title": "Hebbian Robotics - APIs for searching and analyzing Physical AI data at scale", "tagline": "We enable evaluation of data quality without training a robotics model", "body": "# \\*\\*\\*\\*\n\n**TLDR;**\n\n* **Hebbian Robotics builds APIs for searching and analyzing Physical AI data at scale**\n* Our APIs enable data providers to get **on-demand signals and metrics** on the quality of their data, without managing infrastructure\n* We help our customers evaluate data quality without training a robotics model\n\n\n\n**Our Ask**\n\n* Buyers want data they can trust. If you're selling egocentric or robot data, **Let’s Chat ([https://calendly.com/brandon-hebbianrobotics/30min](https://calendly.com/brandon-hebbianrobotics/30min)**) and we'll show you how quality signals and metrics help you sell more at higher prices\n* **Check out Pareto ([https://pareto.hebbianrobotics.com/](https://pareto.hebbianrobotics.com/)),** an example of how our APIs can be used\n\n### **Who We Are**\n\n* We’re Brandon and Kingston, we've trained AI models for industrial cleaning robots and built production systems for camera fleets across the globe\n* [**Brandon Ong**](https://bookface.ycombinator.com/user/388120)'s previous AI startup was backed by Sequoia. He did research on AI for low resource languages that topped Stanford’s LLM benchmark, SEA-HELM, and dropped out of his joint PhD in Robotics at Columbia and NTU\n* [**Kingston Kuan**](https://bookface.ycombinator.com/user/968709) spent years at Jane Street building systems that scaled to millions of messages per second. At Verkada, he worked on their core video surveillance platform that processed large scale multimodal data\n\n### **The Problem**\n\n* **High quality data is the bottleneck for unlocking the next frontier of intelligence in robotics**\n* Buyers want **data they can trust** \n* Data providers don’t have the bandwidth to provide signals and metrics about the quality of their data when they are **focused on** **operational excellence** \n* This means that **model developers are forgoing model improvements due to a lack of high quality data**, impeding progress in robotics\n\n### **Our Solution**\n\n* APIs for searching and analyzing Physical AI data at scale that enables providers to get **on-demand signals and metrics on data quality**\n* Buyers demanding per-collector hour limits? **Prove compliance while we stay ahead of the workarounds**\n* Changed standard operating procedures for collection? **Monitor for quality drift and regression**\n* Suspect duplicate samples? **Verify quickly with a single API call**\n* Our APIs enable **evaluation of data quality without training a robotics model, fast iteration cycles for** high quality data collection, and unlock **model improvements** \n\n### **How to reach us**\n\n* We want to support data providers who believe in collecting high quality data in an ethical way\n* To get started, email [**brandon.ong@hebbianrobotics.com**](mailto:brandon.ong@hebbianrobotics.com)", "url": "https://www.ycombinator.com/launches/SXX-hebbian-robotics-apis-for-searching-and-analyzing-physical-ai-data-at-scale", "votes": 81, "created_at": "2026-08-09T03:00:00.116Z" } ] }, { "id": 33281, "name": "Earendil Robotics", "slug": "earendil-robotics", "former_names": [ "Earendil Robotics Ltd" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/39ec6212427cdf833238bd48eca01f40904ba49f.png", "website": "https://earendil.io", "all_locations": "San Francisco, CA, USA", "long_description": "Earendil Robotics builds cUAS solutions for the modern warfighter, enabling effective air defence at the small-unit level. We deliver spotting, engaging, and destroying capabilities through advanced sensors, autonomous interceptors, and integrated systems.\r\n\r\nOur product lineup covers every stage of the kill chain, complemented by the Drone Warfare Academy that trains soldiers to eliminate enemy drone threats in contested environments.\r\n\r\nBacked by worldwide frontline combat experience, we equip units with combat-proven tools to counter evolving aerial threats.", "one_liner": "Drone swarm defence at a small-unit level", "team_size": 7, "industry": "Industrials", "subindustry": "Industrials -> Defense", "launched_at": 1783628748, "tags": [ "Drones", "Swarm Robotics", "AR" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Defense" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/earendil-robotics", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/earendil-robotics.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/earendil-robotics/", "x_url": "", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/earendil-robotics", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/39ec6212427cdf833238bd48eca01f40904ba49f.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 3723601, "full_name": "James Walton", "title": "Founder/CEO", "founder_bio": "James Walton is Co-founder of Earendil. Former British Army and a UAS specialist with experience in Ukraine, he brings direct operational experience in contested electromagnetic environments and detachment-level drone integration challenges.\n\nPrior to Earendil, he founded and scaled two IT and telecoms companies. His frontline UAS expertise and operational delivery track record position Earendil to close the gap in scalable, NATO-aligned drone warfare training and interception.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/james-walton-79b35339a/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fd50381f327469f83306e46ac417ca633252b142.jpg" }, { "user_id": 3723983, "full_name": "Dan Komorny", "title": "Founder/CTO", "founder_bio": "Dan Komorny is Co-founder of Earendil Robotics. He has a decade of experience in drones and security, including direct operational exposure in Ukraine. He has developed AR heads-up display applications and optical systems, and previously led hardware product development on drone payloads and turrets.\n\nThis combination of frontline insight, product development, and computer vision expertise positions Earendil to deliver advanced UAS and soldier systems.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dan-komorny/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/7fed5227c8ec38cbb843d2210ae2f8e47d2dc39f.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111175, "title": "Earendil | Autonomous Air Defence", "tagline": "Autonomous, combat-proven interceptors for defending against drone swarms.", "body": "**TL;DR:** We build autonomous interceptors that kill drones for less than the cost of the threat. Aeglos is in active operational use and has scored operational intercepts. The team has hundreds of confirmed combat intercepts. Looking for introductions to NATO-aligned air defence and C-UAS leads.\n\n\n\n**The problem**:\n\nAir defence economics are broken. A Patriot interceptor costs $3-7 million and is routinely used against $50,000 drones. Production rates of a few hundred units a year make these systems too scarce and too expensive for the volume of low-cost threats now in the air. They should be reserved for high-value targets.\n\nExactly like the Sherman tank versus the Tiger: it does not matter how good the system is if there are more threats than ammunition.\n\n**What we do**:\n\nAeglos is a high-speed autonomous interceptor designed end-to-end for the mass drone problem: airframe, autonomy software, and operator model. The operator clicks launch, then selects the intercept in software. No stage of flight requires a pilot and no flying experience is needed. One operator can control up to 64 interceptors. Engagement cost stays below the price of the drone being destroyed.\n\nMost C-UAS offerings are a single interceptor airframe. Operators still need skilled pilots, long training pipelines, and constant attention during the engagement. We built the complete system so any operator can achieve consistent intercepts without those requirements. The result is a system that can be fielded at the small-unit and critical-infrastructure level without the specialist crews or multi-million-dollar missiles that current layered air defence relies on.\n\nAeglos is operationally deployed, radar-agnostic, manufactured in Europe, and NDAA-compliant with no Chinese components.\n\n**Why us:**\n\nThe entire team is ex-military with frontline Ukrainian experience, primarily in elite units and air defence. Collectively, we have hundreds of confirmed drone intercepts across multiple systems. We designed Aeglos from that experience - in the same contested Ukrainian airspace where most Western systems have not yet been stress-tested at scale.\n\n[\\@James Walton](https://bookface.ycombinator.com/user/3723601): British Army and UAS expert with frontline Ukrainian experience who has previously run companies.\n\n[\\@Dan Komorny](https://bookface.ycombinator.com/user/3723983): A decade of experience in Drones and Security directly in Ukraine. Serial entrepreneur with multiple startups under his belt.\n\nThe Team: Commandos, Airborne, Special Forces and Air Defense veterans.\n\n**Ask**\n\nIntroductions to air defence and C-UAS leads in NATO-aligned militaries, national acquisition teams, and operators responsible for protecting forward positions and critical infrastructure.\n\nWe are also open to partnerships with complementary companies.\n", "url": "https://www.ycombinator.com/launches/Sv9-earendil-autonomous-air-defence", "votes": 4, "created_at": "2026-08-19T15:45:00.136Z" } ] }, { "id": 33282, "name": "Florin", "slug": "florin", "former_names": [ "Bolten AI" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/48074691217afbc59a4443c6ffa2379a823efda9.png", "website": "https://florin.inc", "all_locations": "San Francisco, CA, USA", "long_description": "Florin is a full-stack insurance carrier where a model does the underwriting. When a submission comes in, Florin reads the ACORD forms, then investigates the risk on its own: satellite imagery, permits, public records, even the insured's Instagram, and catches exposures that never make it onto the application. It then simulates tens of thousands of ways the policy year could go, prices the risk, and returns a bindable quote in under a minute instead of weeks. The team is ex-Google, Tesla, Amazon, and Terrafirma, and the bet is that a carrier built this way wins on loss ratio, speed, and cost at the same time.", "one_liner": "Florin is an AI-native insurance company that runs autonomously", "team_size": 3, "industry": "Fintech", "subindustry": "Fintech -> Insurance", "launched_at": 1782360987, "tags": [ "Fintech", "Insurance", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Fintech", "Insurance" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": true, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/florin", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/florin.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/florinhq/", "x_url": "https://x.com/florin_hq", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/48074691217afbc59a4443c6ffa2379a823efda9.png", "app_video_url": "https://bookface-images.s3.us-west-2.amazonaws.com/bf-public_profile/33340/app_video.mp4?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ASIAQC4NIECAOMAGS7ZB%2F20261020%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20261020T022740Z&X-Amz-Expires=60&X-Amz-Security-Token=IQoJb3JpZ2luX2VjEHEaCXVzLXdlc3QtMiJHMEUCIG5%2FEuPdHreXKRdGLYlyC2AjaYHPccvwucNqtFtDJEpNAiEAs4s5lSDOcogMQrwGhSKvVJXgimMvlIc1KK%2Fa7hojYD4q5QMIOhAAGgwwMDYyMDE4MTEwNzIiDMPa%2BQzgV8CSsX3XwyrCA9yWuwhv7VPDGJqPM%2BDoevlHz47k5lLRdgsGITLyqtSgmLV07z1Y86XDlb4dZMfcWmY5u5FLutDaUy2PMJ92q6vIq41n2jI5tC6SNSed8Qx%2Frdr9Wotko6WcUJWENt0Rs1K4pKZ8ZHCjH22BoWJkWLaTQ8fDKsmgWu0C8NMhnc8i44rKtbZLOvQwKZIlI546tkfyyliSiXPWdZ4XhGxEegateXp4TnM4yWfebpv22hrRHRotmFYX8ANU3B8yiEQL9yrU1y10N8c42JlHR%2FXHZNxaFQqX6eENp2aa2yewya2kPgdiQ3efnk5d0LiuPYxk7ZBa6uOug4fCBPHurDSGZqOlYX9D4h6PAMyUN2kZw2IcUR9QWKM2BmEsGorqokH5Oe34gs0DmnOGrGbhANOh9LKlKDu%2B6Pc4U%2F8L3%2Bmxug0q%2B3JBUGE7Wf6ySj%2B%2BbW8eOteynPrtIEUsD0pcfnSFNzQepsmnoTP5xc7qG%2FEn4MNWO3TviPmcIdg9qyW7bwOWKSyly4BuerWbZWVxIVolnUdqKu2PxDrgnb41QrRtObgKLlbhdehgPlqhIVqZ1cWH9s2uGHoiWlN5FRY8vYYodKzLGjDn%2FdrWBjqlAdXY8WsF5AC2X79ObZ%2FmZD1gjSCNFddCxk1e7RhQVznesCZJDMryJDWqJGSqI7lLEk62aPqOUbquB9Ghba%2BJXI%2FUV%2BHgkJZAPy7wzbnv8vLB8WccujudilaH29xgpqopxweKHuQzlcRMMeq0SZHCo3PizMnbhtLrqB%2BiK9IdfwMp8c15fBnQEe2RAGCsOWYNCrhsPiR6yT5V9698E5KnRPYRfOeFQw%3D%3D&X-Amz-SignedHeaders=host&X-Amz-Signature=ccc17c334dba2900faaea281e2f1ac904074851594b0f3346daf15d3fcdd58f3", "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/50d87df2dfc60c769ee8aa3c3cf2c05dd540d089.png" ], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 3779733, "full_name": "Shaurya Aggarwal", "title": "Founder/CEO", "founder_bio": "Built TPUs in the next-gen TPU team at Google. Before that, worked on greenfield at-rest and in-transit encryption products deployed across 100,000+ hosts. BS in Computer Science from Purdue.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/shauryaagg", "x_url": "https://x.com/shauryaagg", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/eb06d90980c2cf512f38cb7b7c193a45e72e5785.jpg" }, { "user_id": 3776296, "full_name": "Amol Pant", "title": "Founder/CTO", "founder_bio": "Built perception stack at Orchard Robotics, worked on Tesla's exaflop computer for Autopilot, built prototypes for autonomous rovers for NASA Artemis, worked flight simulators for Boeing.\nReach out if you love playing tennis", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/amolpant/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/77b4c26bdf0e4e663044721a47b1a03ff8359262.jpg" }, { "user_id": 3834278, "full_name": "Aydin Sorensen", "title": "Founder", "founder_bio": "7+ year Tech Lead at Amazon AWS & Twitch. Most recently, founding employee at TerraFirma, (vertically integrated construction robotics); scaled realized revenue from $500k to $30m & built foundation for global teleoperation for 50+ vehicle fleet across multiple continents.", "is_active": true, "linkedin_url": "https://linkedin.com/in/peterasorensen", "x_url": "https://x.com/formerlypeter", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/eef2f8e1051ef18328494131e0cc6ed7328ca761.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109455, "title": "Florin: A full-stack insurance company with 0 underwriters", "tagline": "Autonomously places hard-to-price risks in seconds", "body": "**TL;DR:** Florin is a full-stack insurance company with 0 underwriters. Software takes a submission from a raw intake form to a bindable quote in seconds. Instead of matching a business to a class code, it verifies what the business actually does using satellite imagery, permits, public records and social media, then prices it by simulating tens of thousands of ways its next few years could go.\n\n\n\n**What we do and why**\n\nInsurance prices risk with class codes. An underwriter receives an application PDF, matches the business to a row in a rate table, and quotes off that row. This breaks in two ways. Applications are incomplete or wrong, so the price is built on bad inputs. And when a business does not fit any row, either because it is genuinely new (robotics, AI, autonomy) or because the table says the whole class or state is unprofitable, the answer is a declination. That is why carriers are exiting entire states and why a three-year-old industry can barely buy coverage at any price.\n\nFlorin replaces the table with evidence. It ingests the submission directly, no intake team. It builds its own picture of the business from the outside: aerial imagery of the actual premises, permits and public records, even the insured's social media, where it has caught operations that were never disclosed on the application. From that verified picture it runs simulations of the business's possible futures and produces a quote you can bind in seconds. No human touches the file.\n\n![uploaded image](/media/?type=post&id=109455&key=user_uploads/3776296/18477520-2b5a-4473-a727-92daf673768d)\n\n**Team**\n\n\\- Shaurya Aggarwal built silicon on Google's next-generation TPU team\n\n\\- Amol Pant was on Tesla's supercomputer team, designed rovers for NASA, and was early at a Thiel backed robotics startup\n\n\\- Aydin Sorensen was early at Terrafirma, a construction robotics company that recently raised $115M\n\nShaurya and I have been best friends since middle school, and Aydin and I were roommates in college. We all lived together in Seattle for 2 years.\n\n![uploaded image](/media/?type=post&id=109455&key=user_uploads/3776296/4f7bfc6e-40f1-4d8d-9111-97c034db9729)\n\n**Our asks**\n\n1\\. If your business has a risk that regular carriers and brokers keep declining or pricing absurdly, we want to hear from you. Email [shaurya@florin.inc](mailto:shaurya@florin.inc) or grab time at \n\n2\\. If you are a world-class quant, engineer, or underwriter and would rather build the pricing model than read files, we are hiring. Same email.", "url": "https://www.ycombinator.com/launches/STP-florin-a-full-stack-insurance-company-with-0-underwriters", "votes": 194, "created_at": "2026-08-06T14:23:00.075Z" } ] }, { "id": 33283, "name": "Locke", "slug": "locke", "former_names": [ "Overton", "Justinian" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/ef45315b5a753af3ca5decc639d785a08eb897be.png", "website": "https://locke.inc", "all_locations": "San Francisco, CA, USA", "long_description": "Locke is the AI-native firm for public influence. \r\n\r\nWe pair elite lobbyists and political talent with our own AI systems, helping companies lobby policymakers and win government contracts.\r\n\r\nLocke's team comes from The White House, Congress, and the Pentagon.", "one_liner": "The AI-native firm for public influence.", "team_size": 3, "industry": "B2B", "subindustry": "B2B", "launched_at": 1783293519, "tags": [ "GovTech", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/locke", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/locke.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/lockeinc/?viewAsMember=true", "x_url": "https://x.com/lockeai_inc", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/ef45315b5a753af3ca5decc639d785a08eb897be.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 2399943, "full_name": "Parth Badhwar", "title": "Founder/CEO", "founder_bio": "Founder at Locke (YC S26), the AI-native firm for public influence.\n\nPreviously @ Clipbook, The White House, Deloitte's policy team, and the Harris campaign. BA in Political Science from GWU.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/parth-badhwar-8a8037189/", "x_url": "https://x.com/parth_badhwar", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/239987c18eaec74288417c1e7a3b228d44ae015f.jpg" } ], "jobs": [ { "id": 98293, "title": "Member of Policy Staff", "url": "https://www.ycombinator.com/companies/locke/jobs/rTsMRcS-member-of-policy-staff", "location": "Washington, DC, US", "type": "Full-time", "role": "Operations", "role_type": null, "salary_range": "$72K - $200K", "equity_range": "0.10% - 2.00%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 98294, "title": "Member of Technical Staff", "url": "https://www.ycombinator.com/companies/locke/jobs/5HkDZIN-member-of-technical-staff", "location": "San Francisco, CA, US / Washington, DC, US", "type": "Full-time", "role": "Engineering", "role_type": "Full stack", "salary_range": "$120K - $200K", "equity_range": "0.50% - 3.00%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 109419, "title": "Locke - The First AI Government Affairs Firm", "tagline": "Locke is changing the way companies and governments work together.", "body": "[ ](https://youtu.be/AbNIwYUdFBc)\nToday we're introducing Locke, an applied AI company changing the way companies and governments work together. Our first focus: building an AI-native government affairs firm that helps companies monitor regulations, lobby policymakers, and win government contracts.\n\nAmerican governments spend over $10 trillion a year - and it's all up for grabs, reallocated annually through budgets, appropriations, and contracts. But while every other major market has been made accessible, government hasn't. Its front door is guarded by legacy government affairs firms that charge incumbent prices, and their customers get a fragmented product: one firm for the Hill, another for agencies, another for proposals and procurement, with no one accountable for the outcome.\n\nLocke is building the front door government never had. We combine our AI agents with in-house lobbyists who own the entire path - from research to first meeting to funded contract. On the backend, our agents make the path faster than ever - analyzing legislation, tracking stakeholder engagement, and drafting complex contract proposals.\n\nSince quietly launching six weeks ago, we've blown past a $500K annual revenue run-rate serving clients across defense, healthcare, GovTech, and more.\n\nLocke was founded by Parth Badhwar, who came to this problem from both sides of the table: policy roles at the White House and Deloitte, then leading GTM at a startup that sold to dozens of government offices. The team is small, talent-dense, and part of Y Combinator's Summer 2026 batch.\n\nLocke’s mission is bigger than replacing lobbying firms. It's a future where winning government support depends on what you make, not who you know.\n\nIt's day one in that mission.\n\nLearn more at [locke.inc](http://locke.inc) or email [founders@locke.inc](mailto:founders@locke.inc).", "url": "https://www.ycombinator.com/launches/SSp-locke-the-first-ai-government-affairs-firm", "votes": 6, "created_at": "2026-08-06T17:00:00.074Z" } ] }, { "id": 33310, "name": "Petrarch", "slug": "petrarch", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/091a0f9893d5a0f66b65399660fc8041716acd4e.png", "website": "https://petrarch.co/", "all_locations": "San Francisco, CA, USA", "long_description": "Petrarch brings internal company data to frontier labs, starting with manufacturing and industrial companies. We source data including codebases, project files, and payments from bankruptcy courts, then de-identify and prepare the data for model training. \r\n", "one_liner": "Internal industrial company data for frontier labs", "team_size": 3, "industry": "B2B", "subindustry": "B2B", "launched_at": 1782965522, "tags": [ "Reinforcement Learning", "B2B", "Privacy", "Data Engineering", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/petrarch", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/petrarch.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/petrarchinc", "x_url": "https://x.com/petrarchinc?s=11", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/petrarch-e3b1", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/091a0f9893d5a0f66b65399660fc8041716acd4e.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 3655939, "full_name": "Ian Lee", "title": "Founder/CEO", "founder_bio": "Founder at Petrarch. Previously Fellow at Meridian AI building training data and GTM infrastructure for AI-native Excel in financial services. Technical Advisor at Scale AI, working on quality filtering and annotation for agentic assistants and reasoning models. BA in Applied Mathematics and Psychology from Harvard.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/-ianlee", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/bceb6907c6226831472da7b1d7688785c1fb3c17.jpg" }, { "user_id": 2670953, "full_name": "Samuel Hahn", "title": "Founder/COO", "founder_bio": "Founder at Petrarch. Previously led GTM at Magier AI (Techstars '24), building agentic PII-redaction and data-compliance products. Venture analyst at Mainstreet sourcing pre-seed fintech and deeptech. BA in Economics and History from Harvard.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/samuelhahn9/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c4c5f0acb76bd4a03ee95588c1e5324406ca2dae.jpg" }, { "user_id": 3594340, "full_name": "Sudhish Swain", "title": "Founder/CTO", "founder_bio": "Founder at Petrarch. Previously built computer vision software for medical equipment identification deployed in California hospitals as Head of Software Engineering at Clean Sweep Group. Research intern at Brigham and Women's Hospital analyzing spatial transcriptomic cancer datasets. BA Integrative Biology at Harvard.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/sudhish-swain/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/098958191b28eaf5108b38cb0d2768964c348d55.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 105007, "title": "Petrarch — Unlocking enterprise data for tomorrow's economy", "tagline": "The market for specialized data locked up in enterprise companies.", "body": "# The market for specialized data locked up in enterprise companies.\n\n[Petrarch](http://petrarch.co) is unlocking enterprise data for tomorrow's economy.\n\n\n\nFrontier labs scraped the internet dry. Then, labelers contracted subject matter experts to hand-make data. Now demand has shifted towards the bespoke data locked up within industries like construction, manufacturing, and logistics.\n\nPetrarch targets companies with valuable **operational** **data** they haven't monetized, from **legacy** **businesses** and **startups** to companies undergoing **bankruptcy** and **pivots**.\n\nPetrarch **sources** suppliers, **cleans** and **restructures** the datasets, certifies **legal clearance**, and resells to **applied AI startups and labs.** They have already transacted construction plans, autonomous driving recordings, and codebases.\n\nWith experience at Scale and Meridian AI, founders Ian Lee, Samuel Hahn, and Sudhish Swain dropped out of Harvard to unlock the next wave of training data.\n\nIf you need proprietary datasets or think you have something valuable on your hands \\\n(or recently pivoted), reach out at [founders@petrarch.co](mailto:founders@petrarch.co) or visit us at [Petrarch.co](http://Petrarch.co)!", "url": "https://www.ycombinator.com/launches/RJf-petrarch-unlocking-enterprise-data-for-tomorrow-s-economy", "votes": 11, "created_at": "2026-07-13T16:11:02.837Z" } ] }, { "id": 33311, "name": "SpaceFlow Technologies, Inc.", "slug": "spaceflow-technologies-inc", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/161339dc99cea591f4bd07fbb627aaa0b7771a88.png", "website": "https://spaceflow.tech", "all_locations": "San Francisco, CA, USA", "long_description": "SpaceFlow is an AI-native procurement services company for businesses that make, move and serve physical goods. We help procurement teams bring more spend under management, execute source-to-contract work and turn identified savings into realized savings without adding headcount or replacing existing systems.", "one_liner": "AI-native procurement services for the physical economy", "team_size": 4, "industry": "B2B", "subindustry": "B2B -> Supply Chain and Logistics", "launched_at": 1785644557, "tags": [ "Fintech", "Logistics", "Manufacturing", "Supply Chain" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Supply Chain and Logistics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/spaceflow-technologies-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/spaceflow-technologies-inc.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/spaceflow-yc-s26/", "x_url": "https://x.com/usespaceflow", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/spaceflow-de69", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/161339dc99cea591f4bd07fbb627aaa0b7771a88.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/3c6238c718981e5757df84d1247509ab56f2c870.png" ], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 1732874, "full_name": "Emre Isik", "title": "Founder/CEO", "founder_bio": "Founder and CEO of SpaceFlow Technologies. Previously researched computational linguistics at Cambridge under Prof. Anna Korhonen and developed system identification algorithms for motor control at UCLA's Institute of Pure and Applied Mathematics. BA in Computer Engineering from University of Cambridge. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/emre-isik-6a88aa193/", "x_url": "https://x.com/Emrre_isik", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/28d05ebaf53739897d5635366d7199420e898744.jpg" }, { "user_id": 3610149, "full_name": "Hakan Enes Aksu", "title": "Founder", "founder_bio": "Co-founder and CTO of SpaceFlow Technologies. Previously build crypto-ad automation software in high school, AI models for sports analysis and built a campus social platform. Dropout", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/hakan-enes-aksu/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a232238a981347ef4a7cd6905b71a4afe540049a.jpg" }, { "user_id": 2475947, "full_name": "Ali Eren Aytekin", "title": "Founder", "founder_bio": "Co-Founder of SpaceFlow Technologies. Previously developed deep learning models for genetic perturbation analysis, earning runner-up at the Microsoft Imagine Cup. Holds a patent for a compressed-air energy storage system. Engineering at Istanbul Technical University, dropout.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ali-eren-aytekin-6305a41aa/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/09e1fd106720cfbfd5b526bcaa939f3b64eaef64.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109415, "title": "SpaceFlow: AI-native procurement services for the physical economy", "tagline": "An extension of your team handling sourcing, planning and contracting. ", "body": "**TL;DR:** SpaceFlow is an AI-native procurement services company for businesses that make, move and serve physical goods. We help procurement teams bring more spend under management, execute source-to-contract work and turn identified savings into realized savings without adding headcount or replacing existing systems.\n\n\n\n### The problem\n\nProcurement must absorb changing demand, supplier issues, and growing volume without increasing cost or risking continuity. Every existing option fails on the same axis:\n\n* **Internal teams:** expertise and judgment, fixed capacity. It takes years to build and when someone leaves, the tacit knowledge that keeps operations running is lost.\n* **SaaS:** automation and visibility without execution. The customer still staffs the operation, supervises it, handles exceptions, and owns the result.\n* **Outsourcing:** capacity, expertise, and accountability, permanently labor-linked. More volume, suppliers, and locations require more service-delivery headcount.\n\nThe unmet need is operating capacity that **takes responsibility for completing the work** while becoming faster, more consistent, and more profitable as volume grows.\n\n## The insight\n\n* **Software can recommend actions; it cannot assume commercial authority.** The customer keeps execution, exceptions, and supplier outcomes. Whoever owns the exceptions owns the expertise, and expertise is the entire product in procurement.\n* **Incumbents cannot follow.** Consulting and outsourcing firms store process knowledge in employees and grow revenue through staff hours. Automation directly cannibalizes that model.\n* **The economics are breakable.** Traditional services sit below software economics: Accenture at Rule-of-22.6, Genpact at 24.1. An AI-native operator can reach **Rule-of-60**.\n\n## What we built\n\n**The proprietary platform SpaceFlow uses to operate procurement for customers.**\n\n* A **process-mapping engine** converts interviews, SOPs, approval rules, and existing system configurations into executable workflows. A supply-chain consultancy estimates it cuts customer launch from **26 months to 4**, a 6.5x jump in deployment capacity.\n* It **executes procurement end to end**, from demand and production planning through sourcing, contracting, purchase orders, exception resolution, and reconciliation, across the customer's own systems.\n* Operating across companies in the same vertical turns cross-company experience into **proprietary process and supplier intelligence**, while pooled demand creates **group-purchasing power**: better execution, prices, resilience, and terms for every customer.\n* The same platform **monitors service levels, resolves exceptions, and remains accountable** for completing the operation. Per-action autonomy limits, approval queues, and an immutable audit log make every task provable.\n\n## Proof so far\n\n* **Live:** $400M a year of supplier spend under management, zero systems migrated. Multiple six figure contracts signed.\n* **Demand for delegation:** our first customer cancelled 3 planned procurement hires, the CSCO invested his own money, and two chains are signing to progressively outsource purchasing, one planning a procurement team 90% smaller by 2027.\n* **Results:** 3x faster RFQ cycles, off-contract spend down 42% at one customer, $120K a year saved at another.\n\n## Where this goes\n\nSpaceFlow will replace legacy service providers across the physical economy. Traditional service firms grow by hiring; SpaceFlow grows output faster than headcount, targeting **$1B in revenue with headcount capped at 100**. \n\n## What's in it for you?\n\n1. **You run a physical-goods company with $100M+ supplier spend:** send us 12 months of PO history and we'll map one lane and size the savings, free. [spaceflow.tech](http://spaceflow.tech) · [emre@spaceflow.tech](mailto:emre@spaceflow.tech)\n2. **You know an operator like that:** intros are gold. \n\n### **About the team:**\n\n**Emre Isik (CEO)**  studied computer engineering at the University of Cambridge and dropped the master's to build SpaceFlow full time. Before that he researched embedded AI for motor control at UCLA's Institute for Pure and Applied Mathematics, published a machine-learning paper on reasoning with small data, and trained with Turkey's international informatics olympiad reserve team at 15. \n\n**Ali Eren Aytekin (co-founder)** left engineering at Istanbul Technical University to build SpaceFlow.  He was Microsoft Imagine Cup EMEA runner-up for a genetic-signal classifier that hit 98.3% accuracy, and co-authored the published paper on low-data reasoning. [LinkedIn](https://www.linkedin.com/in/ali-eren-aytekin-6305a41aa/) . He talked his way into an invite-only AI showcase in Istanbul, came out with a procurement problem from the CEO of a UK medical manufacturer, and shipped the solution with the team in two weeks. That deal became the learning SpaceFlow grew from.\n\n**Hakan Enes Aksu (co-founder)** left information engineering at Boğaziçi University to build SpaceFlow. He has been shipping commercial software since 16, when his Reddit ad-automation service earned $40K in revenue, and he later built a tender-matching engine that reads public tenders against a company's product catalog and sold it to a corporate customer for five figures. He scored in the top 500 out of roughly 3 million students in Turkey's national university entrance exam. [LinkedIn](https://www.linkedin.com/in/hakan-enes-aksu/)", "url": "https://www.ycombinator.com/launches/SSl-spaceflow-ai-native-procurement-services-for-the-physical-economy", "votes": 37, "created_at": "2026-08-06T15:30:00.856Z" } ] }, { "id": 33312, "name": "Ethos Space Resources", "slug": "ethos-space-resources", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/26c677c7621160c55e8011d6851971173fe4f8e9.png", "website": "https://ethos-space.com", "all_locations": "Los Angeles, CA, USA", "long_description": "Ethos makes silicon on the Moon to power terawatt scale data centers. The Moon's surface is 21% silicon. We melt and electrolyze the lunar surface to make raw materials for solar panels. With solar panels made from the Moon, the constraint that limits data center scaling and everything else in space is gone.", "one_liner": "We make silicon on the Moon, to power everything in space.", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Aviation and Space", "launched_at": 1787528961, "tags": [ "Hard Tech", "Exascale Computing", "Energy", "Aerospace", "Industrial" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Aviation and Space" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": true, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/ethos-space-resources", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/ethos-space-resources.json", "year_founded": 2022, "city": "Los Angeles", "country": "US", "linkedin_url": "https://www.linkedin.com/company/ethos-space/", "x_url": "https://x.com/ethosspace", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/26c677c7621160c55e8011d6851971173fe4f8e9.png", "app_video_url": "https://bookface-images.s3.us-west-2.amazonaws.com/bf-public_profile/33370/app_video.mp4?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=ASIAQC4NIECAOTKCHR7A%2F20261020%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Date=20261020T022740Z&X-Amz-Expires=60&X-Amz-Security-Token=IQoJb3JpZ2luX2VjEHEaCXVzLXdlc3QtMiJHMEUCIQDeV7cQCYWa%2ByyakJaw7aH8e0HNWEyd2tO1Fn9KHjIpkgIgJPUveNomWrSf4Oq5DPuVDY7LKeaNJPvi5F2sbSx87Hoq5QMIOhAAGgwwMDYyMDE4MTEwNzIiDJ7%2FkqUU8p5mIeDC5CrCA7F4UggTYlVYiff4mYkoYiIpXa9r%2FKkUhcKwMNwtYpcuNTEBtH5xi9Hzkl3B%2FOYmUn%2FIacrdemsB9KrtOuRo%2FccaY%2Bs%2FSuIzime%2BcYuI7%2B%2Fg2nldjcm7YPCf32sIR8EflXc2916%2FDRMbztsRKDiQDL1ZKbK6YPHJppiQyVyRrrcCFD0X7tEf0D7QgK2DIpKnu8zGLzN4oAHn6xfHIgQnynrupcJbc9cga5AURtsRl7TfSM6XaGTtp1fNogyhUKX9f%2FYIL2pg03nlhLpRpDuqHpJ9qbbVsf%2FvboDHQ4IRRkiCaWH6GZwuTm%2FaXzTPIrAlVKH5%2F0RNaaExh%2Byw4VhHEEsSkHSs4mw%2BmNAKnW2FnWuldPRw5kn3j05dqhv5fkQxdp74JTYQe5wex1YN5PCok147bSKEuieYlIgH%2FiP9NY3Ym0ZRUWuo0zL40XWv9X3L1tLg%2BYWY0fCdhDKHGgTuesmbVajzEhu%2Bg9iWLptoK63cZ4UfO8oOd6sOc09XF6lBy%2BmD%2BSVmOfDWUOzz5ELD16slDTUwml0MZsttBoLDzldUc0olkl5nCAl23UqraiUXeaeW3yLOvJn%2BAD%2B%2FM6PvmZF0VzDn%2FdrWBjqlAVRsuTd0UsfPEqyKlBgBcSJTbWkHvSj10wzTCIqvOoZ3bz1TWza43iHLTCs7u%2FHmvoKE0qG5FTOsMYhLZb5XFdRretuo%2F1BjdoZt%2Fr0dHWGT4LLkmfYrkFclUq7u0pvMr6%2FR4I41ZYdCD3cWpGAcOrCQxKJukkkU%2Fx4O0IOnvZKvXzK4OISHe4XPgWQrnAJcSBOFX%2BXRwhpFbVXOOC%2BtEw%2BR3JTElw%3D%3D&X-Amz-SignedHeaders=host&X-Amz-Signature=60b56f24c8f17b5459a5b8412a9a835e1539c0cf6f5658b9bcb033320edc6ca0", "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/252852e7026776dd179fcebea3781488d68578fc.png" ], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 143296, "full_name": "Ross Centers", "title": "Founder/CEO", "founder_bio": "Lunar resource geologist. Founder of Planetary Sunshade Institute. Kardashev scaling since 2015.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/centers/", "x_url": "https://x.com/rosscenters", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/a7da888d8bb4879eb5ed964886a037aa38c0859a.jpg" }, { "user_id": 3684585, "full_name": "Brogan BamBrogan", "title": "Founder/CTO", "founder_bio": "Employee #23 at SpaceX. Designed and built the 2nd stage engine for Falcon 1 as well as many elements of the Dragon spacecraft, including the primary heat shield, Draco thrusters, and more. Later co-founder & CTO at Hyperloop One and co-founder & CEO at Arrivo. Earlier engineering roles at Northrop Grumman and Chrysler Motors.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/brogan-bambrogan/", "x_url": "https://x.com/BamBrogan", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/589e273af2b0a3ff1590082a9389e7e3a73ecb05.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112089, "title": "Ethos Space Resources", "tagline": "Silicon on the Moon for Terawatts in Space", "body": "**TL;DR**\n\n**Team**\n\nBrogan was employee #23 at SpaceX. He built the first Falcon and Dragon. Previously, he founded Hyperloop and Arrivo.\n\nRoss is a lunar geologist from the Colorado School of Mines. He previously founded the Planetary Sunshade Institute.\n\n**Problem**\n\nThere is infinite power in space. But rockets can’t carry enough solar panels. Rockets are the constraint.\n\n\n\n**Solution**\n\nPower in space must be made in space. Self replicating power and mass relieves the rocket constraint.\n\nThe Moon is made of Silicon (and other useful atoms). Silicon makes solar panels (and compute). Solar panels make energy, to refine more silicon and power more compute.\n\n\n\n**Traction**\n\n* Completed NASA contract\n* $1B LOIs from multiple customers\n* Going to the Moon in 2028\n\n**Why now**\n\nCapital needs from Earth will diminish as self-replicating power and mass grows on the Moon. Moving up the Kardashev scale will be a matter of wattage and tonnage, not dollars. If you’re not early to self-replication, you’re too late.", "url": "https://www.ycombinator.com/launches/T9t-ethos-space-resources", "votes": 17, "created_at": "2026-08-25T15:00:00.072Z" } ] }, { "id": 33313, "name": "SubVysion", "slug": "subvysion", "former_names": [ "SubVysion, Inc." ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/0df819d609df4c8e15ebdc21bbc1f518d56e52d8.png", "website": "https://subvysion.com", "all_locations": "San Francisco, CA, USA", "long_description": "We build centimeter-accurate 3D maps of underground utilities to help construction contractors dig safer and faster. With underground intelligence, contractors can avoid dangerous utility strikes and costly project delays.", "one_liner": "Autonomous rovers to make the ‘Google Maps’ of underground utilities", "team_size": 2, "industry": "Real Estate and Construction", "subindustry": "Real Estate and Construction -> Construction", "launched_at": 1783960374, "tags": [ "Hard Tech", "Computer Vision", "Real Estate", "Construction", "Infrastructure" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Real Estate and Construction", "Construction" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/subvysion", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/subvysion.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/subvysion/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/0df819d609df4c8e15ebdc21bbc1f518d56e52d8.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 1363066, "full_name": "Maxwell Stone", "title": "Founder/CEO", "founder_bio": "Founder at SubVysion. Previously Machine Learning Engineer at Roboflow, where he built on-device inference SDKs deployed in 10,000+ pharmacies for a Fortune 100 healthcare company and developed ML solutions for enterprise customers including railyard mapping and defect detection. MSE and BSE candidate in Computer Science from UPenn.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/maxwell-stone-2b5b6b1a0?utm_source=share&utm_campaign=share_via&utm_content=profile&utm_medium=ios_app", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/534fdc481c3e4f0ac1b1ba8eb1dbfcf6b3c19b39.jpg" }, { "user_id": 3411443, "full_name": "Wenson Tang", "title": "Founder/COO", "founder_bio": "Co-founder and COO at SubVysion, Inc. Computer Science and Consumer Psychology from UPenn. Previously Software Developer at CentralSquare Technologies, working cross-jurisdiction emergency dispatch allowing first responders to reach emergencies three minutes faster.\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/wenson-tang", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6138c520a0b71cd0f533baf3b0cd765d9ffc1ddd.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111406, "title": "SubVysion: Preventing $150B is wasted construction spending", "tagline": "Autonomous rovers building the 'Google Maps' of subsurface infrastructure", "body": "\n\nConstruction has a mapping problem. Of the $2.2T in annual US construction spending, $150B is wasted due to a lack of subsurface knowledge. All construction starts from the ground up, and more often than not, unexpected assets underneath the surface will hold a construction project back. Whether it's an unknown utility line, unexpected rock layer, or old foundation that wasn’t on any design map, the costs of delays from underground conflicts quickly add up. \n\nTake utility strikes, when an excavator hits an active utility line. For every dollar spent on the explicit repair cost of that utility line itself, the contractor loses an additional $30 per repair-dollar in project delays, insurance claims, idle work crews, and increased insurance premiums. For a hyperscale project such as data center construction, a single day of delay can cost the cloud-provider up to $14 million dollars for that single day.\n\nIn metropolitan areas, maps are incomplete at best and none create a unified picture of critical infrastructure, making excavation especially dangerous when natural gas or electric lines are broken.\n\nEnter SubVysion. We’ve built a fleet of autonomous rovers carrying ground-penetrating radar and other sensors that create a centimeter-accurate map of subsurface infrastructure. Our hardware is the first and the best to map the entirety of US infrastructure. Using our rovers, we comprehensively map construction sites, city streets, and even larger indoor buildings, at 100x better granularity and 10x the speed (with a single rover) compared to our competitors. We process vast amounts of data using machine learning and agents, and output that data into an easy-to-understand 3D scene. \n\nWho are we? Maxwell Stone was an early ML engineer at Roboflow (YC S20), and Wenson Tang built the world’s smallest and cheapest LiDAR-capable drone for adverse environments. We both studied computer science at the University of Pennsylvania. Together, we’re building SubVysion to map the entirety of America’s subsurface infrastructure.\n\nWe’re currently on job sites in Philadelphia and Texas, helping our first market of utility contractors save hundreds of thousands of dollars every year. We’re soon to expand our mapping services to municipalities and energy providers as we scale up our rover fleet and capacity!\n\nIf you’re in the energy, infrastructure, or hyperscale construction world, we’d love to hear from you at [founders@subvysion.com](mailto:founders@subvysion.com)!", "url": "https://www.ycombinator.com/launches/Sys-subvysion-preventing-150b-is-wasted-construction-spending", "votes": 7, "created_at": "2026-08-19T23:30:02.157Z" } ] }, { "id": 33314, "name": "Avoca Systems", "slug": "avoca-systems", "former_names": [ "Avoca" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/555346995ae2b6c3163f5ec605c10e5d27db2e8d.png", "website": "https://avocasystems.com", "all_locations": "San Francisco, CA, USA", "long_description": "Avoca is the AI patient access platform built for radiology. It integrates with a network’s existing systems and operational knowledge to automate calls, online scheduling, faxes, e-referrals and patient communications, from referral through to results and follow-up. Unlike generic scheduling tools, Avoca understands the clinical, operational and resource constraints behind every radiology booking, helping providers capture more demand, reduce administrative work and fill more scan capacity.", "one_liner": "The AI operating system for radiology networks", "team_size": 6, "industry": "Healthcare", "subindustry": "Healthcare", "launched_at": 1784843716, "tags": [ "Healthcare", "Enterprise Software", "Healthcare IT" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/avoca-systems", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/avoca-systems.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/avocasystems/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/555346995ae2b6c3163f5ec605c10e5d27db2e8d.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 2570497, "full_name": "Max Shand", "title": "Founder/CEO", "founder_bio": "Co-founder of Avoca, the operating system for outpatient radiology networks. Previously, exited founder of Serenade, innovative music merchandise platform for artists like Oasis, Yungblud and Lewis Capaldi. Formerly, first employee at Afterpay, helping scale from idea to $20BN exit to Block. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/max-shand-700a3ab5/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/dd4429ce745d577782dcef4fdc9b7faadb2547ee.jpg" }, { "user_id": 3763127, "full_name": "Campbell Mercer-Butcher", "title": "Founder/CTO", "founder_bio": "Co-founder at Avoca Systems (YC S26). Principal engineer at Honey Insurance, Macquarie Bank, and Ailo. Built Macquarie's robo-advice platform, used by 1,000+ financial advisors, and automated claims processing at Honey, replacing manual review with instant decisions.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/campbell-mercer-butcher", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4e127b66cb6e5c86db96d2e2ba0a173654e316e6.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107202, "title": "Avoca Systems: AI-powered radiology scheduling and communications", "tagline": "Automating radiology scheduling and patient communications", "body": "Hey everyone 👋\n\nWe’re Max Shand and Campbell Mercer, co-founders of Avoca Systems.\n\nTL;DR: Avoca automates patient scheduling and communications for radiology groups. Our radiology-specific AI handles calls, online scheduling, faxes and e-referrals, guiding patients from referral and booking through to results and follow-up. We’re already live across more than 250 clinics, including networks operating 150+ locations.\n\n\n\n**What it does**\n\nRadiology scheduling is unusually complex. A valid appointment can depend on the modality, body region, contrast requirements, patient history, preparation instructions, equipment availability, staff qualifications and referring-provider details.\n\nThese rules are fragmented across legacy systems, spreadsheets and the tacit knowledge of experienced receptionists. Generic voice and scheduling tools struggle to bring them together.\n\nAvoca integrates with a radiology network’s existing systems and turns its workflows and operational knowledge into one automation layer:\n\n* **AI calls:** Answers inbound calls and runs outbound campaigns for bookings, rescheduling, cancellations, preparation questions and general enquiries.\n* **Online scheduling:** Lets patients upload their referral, complete the necessary clinical and safety questions, and find an appropriate appointment in real time.\n* **Fax and e-referral automation:** Ingests and interprets referrals, identifies missing information and contacts patients to complete their booking.\n* **Operational intelligence:** Encodes each network’s booking rules, rosters, staff capabilities, preferences and tacit knowledge so Avoca can safely fill available scanner capacity.\n* **Patient communications:** Guides patients through preparation, intake, reminders, results and follow-up using voice, text and email.\n\n**Why this matters**\n\nUp to 30% of calls into radiology centres go unanswered as patients wait on hold. Referrals arrive through phones, websites, faxes and electronic systems, leaving staff to manually interpret documents, contact patients and coordinate appointments across multiple systems.\n\nAt the same time, radiology providers face staff shortages, rising operating costs and reimbursement pressure. Missed calls and unprocessed referrals mean patients wait longer while valuable scanner capacity goes unfilled.\n\nAvoca helps networks capture more patient demand, reduce administrative work and complete more scans using their existing clinics, equipment and clinical systems.\n\n**Traction**\n\n* Partnered with 250+ clinics, including several of Australia’s largest radiology networks\n* Reduced call abandonment from approximately 30% to 2%\n* Patient satisfaction averaging 4.6/5\n* More than 80% booking conversion for eligible patient calls in major deployments\n* Integrated with leading radiology and healthcare systems, including Visage, Karisma, Comrad, Epic, Oracle Health and eClinicalWorks\n* Now expanding into the United States\n\n**Why we’re building this**\n\nI (Max) was the first employee at Afterpay and later founded and sold Serenade. Cam was previously a principal engineer at Honey Insurance and Macquarie. Across those experiences, we saw how much better an industry can operate when complex workflows are turned into simple, intuitive products.\n\nHealthcare is a chaotic mess of work that falls between systems: calls that go unanswered, referrals that sit in fax queues and operational knowledge that exists only in the heads of experienced staff.\n\nWe started with radiology because the scheduling complexity is exceptionally high and the consolidation of radiology clinics into large groups allowed us to focus our GTM efforts on just a few influential stakeholders who acutely feel the pain we’re solving for.\n\n**Our big** **ask**\n\n* Any introduction to a radiology network or health system chief executive (CEO, CIO, COO, CDO, CRO - even the new brand of CAIO - Chief AI Officer) would be hugely appreciated. We’re new to the US and eager to meet leaders tackling the patient access and scheduling challenges we’re solving.\n\nFinally, if you operate a radiology network - or know someone who does - we’d love to hear from you. Please reply here or reach us at [hello@avocasystems.com](mailto:hello@avocasystems.com).\n\n![uploaded image](/media/?type=post&id=107202&key=user_uploads/2570497/a75203b9-8ce9-44cc-a0dd-b456dafd0525)\n\n", "url": "https://www.ycombinator.com/launches/Rt4-avoca-systems-ai-powered-radiology-scheduling-and-communications", "votes": 4, "created_at": "2026-07-28T22:04:43.714Z" } ] }, { "id": 33316, "name": "Tracer", "slug": "tracer", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/7ad5cca73c945be7e986c41180c71d7ef7a7b19c.png", "website": "https://tracerml.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Tracer is an AI research lab building more capable and efficient AI through model coordination. We research adaptive inference, ensembling, and systems in which multiple open-source models work together to produce capabilities beyond any single model.\n\nOur work spans reasoning, code, research, and agentic systems, with the goal of making advanced intelligence more accessible and economical.", "one_liner": "Combining open-source AI models for better answers at lower cost", "team_size": 1, "industry": "B2B", "subindustry": "B2B", "launched_at": 1785291116, "tags": [ "Machine Learning", "Open Source", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/tracer", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/tracer.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/116774192/", "x_url": "https://x.com/tracerml", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/7ad5cca73c945be7e986c41180c71d7ef7a7b19c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 1217984, "full_name": "Adam Rida", "title": "Founder/CEO", "founder_bio": "I’m the founder of Tracer (S26), an AI research lab building more capable and efficient intelligence by coordinating open-source models.\n\nBefore Tracer, I built DeepRecall solo to €100k ARR in three months, scanning 100k+ marketplace products per day for product-safety risks.\n\nMy background is in applied machine learning and explainable AI. I was an AI PhD candidate at Sorbonne University, CNRS, and AXA, and previously worked at Société Générale, Rebellion Research, Qantev, and MFG Labs.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/adam-rida-581296142/", "x_url": "https://x.com/_adamr_1", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2fea71058220568dee5bc25f3d49796d0da28cad.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108361, "title": "Tracer: Fable-level AI at 1/3 the cost using open-weight models", "tagline": "Adaptive AI systems that combine open-weight models for better answers at lower cost.", "body": "Hi! I’m Adam, founder of Tracer.\n\nWe built **Echo**, an adaptive AI system built entirely from open-weight models. Instead of making you choose between models, Echo decides how much compute each request needs, which models should work on it, and how their answers should be combined.\n\n**Across our current eval mix, Echo reaches Fable-level performance at roughly 1/3 the inference cost.**\n\nToday, Echo is strongest on math, science, and the coding benchmarks we currently cover. Multilingual and broader knowledge performance still has room to improve, while longer-horizon coding and agentic evaluations are underway.\n\nEcho is continuously updated as we expand the model pool and improve ensembling and cross-model coordination.\n\n\n\n## **Why we built Echo**\n\nThere are now dozens of strong models, each with different strengths, weaknesses, and prices.\n\nWe started with a simple experiment: what if you had a perfect oracle that always knew which model would perform best for each request?\n\nThe oracle was substantially stronger than any individual model in the pool.\n\nEcho is our attempt to recover that advantage without knowing the answer beforehand.\n\nInstead of picking one model for everything, Echo goes beyond routing and dynamically allocates compute across a pool of open-weight models behind a single interface.\n\n## **The result**\n\nWe evaluate Echo against frontier models across reasoning, math, coding, multilingual and other benchmarks.\n\nWe measure both **quality and the actual model inference cost used to produce the answers.**\n\n![uploaded image](/media/?type=post&id=108361&key=user_uploads/1217984/867cc6aa-16b8-4330-b497-c33d78e62b24)\n\nOn MATH-500, Echo reached **98.6% at $4.58**, compared with Claude Fable at **99.8% at $12.64**.\n\nOn LiveCodeBench, Echo reached **92.8% at $5.35**, compared with Claude Fable at **92.0% at $8.71**.\n\nThese results show where Echo is already strongest. LiveCodeBench is encouraging, but it does not capture the full difficulty of repository-level and agentic coding. We are currently running longer-horizon evaluations where Echo must plan, use tools, modify code, and recover from errors across multiple steps.\n\nThose results are visible in the Eval Observatory and are active areas of improvement.\n\nWe publish the underlying evaluations, including individual answers, costs, wins, and losses:\n\n**→[ Explore the Echo Eval Observatory](https://echo.tracerml.ai/eval?utm_source=chatgpt.com)**\n\n## **From open source to Echo**\n\nWe first explored this idea through Tracer's open-source work: learning from real AI workloads where expensive model calls are actually necessary and where cheaper paths are enough.\n\nThe project has now crossed **1,000 GitHub stars**, with developers using and experimenting with the approach in their own AI pipelines.\n\nEcho is the next step. Instead of optimizing individual calls with a routing layer, we're building one adaptive system that decides how intelligence should be allocated end to end.\n\n**→ [GitHub](http://github.com/adrida/tracer)**\n\n## **Every company’s AI traffic has a shape**\n\nOur public evaluations test Echo across a broad mix of general tasks. Production workloads are different.\n\nEvery company has its own traffic distribution and quality requirements. Some requests genuinely need expensive frontier-level intelligence. Many don’t.\n\nFor teams that work with us, Echo can be configured to learn the structure of their workload and identify where expensive intelligence is useful and where cheaper paths are enough.\n\n**Public Echo does not train on users’ private traces. Workload-specific learning is enabled only for teams that explicitly work with Tracer to configure it.**\n\nWe set up this optimization with each team around its workload, quality bar, and data requirements.\n\nThe goal is simple: **an AI system that continuously learns how to spend the minimum amount of inference required to maintain the quality you need.**\n\nThat means **\\~3x cheaper is not necessarily the ceiling for a real production workload.**\n\n## **Try Echo now**\n\nEcho is publicly available through a chat UI and an **OpenAI-compatible API**.\n\nYou can try it directly, connect an existing application by changing the base URL, or plug Echo into OpenCode and other agent harnesses.\n\nEcho is continuously updated as we improve model selection and ensembling.\n\n**→[ Try Echo now](https://echo.tracerml.ai)**\n\n## **We’re looking for teams with large inference workloads**\n\nIf your company is running meaningful LLM inference in production, we’d like to test Echo against your workload.\n\nWe’ll map the quality-cost frontier on your traffic and determine how much of your inference actually requires expensive intelligence.\n\nTeams can then work with us to activate private, workload-specific optimization.\n\n→ [**Optimize Echo for your workload**](https://cal.com/adam-rida-tracer/tracer-first-call)\n\n## **About Tracer**\n\nI’m **Adam Rida**, founder of Tracer. I was previously a PhD candidate at **Sorbonne University/CNRS**, where I conducted and published research in ML interpretability. Before founding Tracer, I built applied ML systems across finance, insurance, and supply chain, then grew [DeepRecall](http://deeprecall.io/) to \\~**€100k ARR as a solo founder in less than 3 months**.\n\nI’ve spent most of my career studying how models make decisions and turning that research into production systems.", "url": "https://www.ycombinator.com/launches/SBl-tracer-fable-level-ai-at-1-3-the-cost-using-open-weight-models", "votes": 18, "created_at": "2026-07-31T21:01:56.160Z" } ] }, { "id": 33317, "name": "Wondering", "slug": "wondering", "former_names": [ "Mouseover AI", "Mouseover" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/188e05122015c2a6e85c91c5145eeef637482a9b.png", "website": "https://wondering.app", "all_locations": "San Francisco, CA, USA", "long_description": "Wondering turns anything you want to learn, from LLMs to quantum mechanics to product design, into a path of three-minute lessons with visuals, podcasts, and interactive exercises. \r\n\r\nEach learning path and its content are personalized to your background, how you learn best, and how deep you want to go.\r\n\r\nIt’s the most delightful way to understand complex topics and turn information into knowledge you can remember and apply. You don’t just passively read. You practice, make connections, and form your own insights.\r\n\r\nStart learning anything for free at https://wondering.app/download", "one_liner": "Duolingo for learning anything", "team_size": 2, "industry": "Consumer", "subindustry": "Consumer", "launched_at": 1785135586, "tags": [ "AI-Enhanced Learning", "Education", "Consumer", "AI", "AI Assistant" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Consumer" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/wondering", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/wondering.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/wondering-app", "x_url": "https://x.com/wonderingapp", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/188e05122015c2a6e85c91c5145eeef637482a9b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 1422401, "full_name": "Cheng-Wei Hu", "title": "Founder/CEO", "founder_bio": "Co-founder at Wondering. Previously at NotebookLM. Interested in the art and science of learning.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/cwhu/", "x_url": "https://x.com/HcwXd", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/360a2c8064324de55565afd3eb1eff18ca05d50a.jpg" }, { "user_id": 3675432, "full_name": "Angelica Kosasih", "title": "Founder/CTO", "founder_bio": "Co-Founder of Wondering\nPrev Meta", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/angelica-kosasih/", "x_url": "https://x.com/aka_kosasih", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/983cdbc8a960ece55f686ee0dfc3feb02501dbfd.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107951, "title": "Wondering - Duolingo for learning anything", "tagline": "Break down any complex topic into a personalized path of short interactive lessons", "body": "Wondering is the most delightful way to understand complex topics and turn information into knowledge you can remember and apply.\n\nTell it what you want to learn, from[ LLMs](https://wondering.app/learn/attention-is-all-you-need-zp2vvu) to[ Analytic Philosophy](https://wondering.app/learn/tractatus-logico-philosophicus-qfmwam), and it creates a structured path of short lessons with visuals, podcasts, and interactive exercises, personalized to what you already know and how deep you want to go.\n\nEach lesson surfaces the most important ideas and lets you customize how you explore them, like a thoughtful tutor beside you.\n\nCheck out our[ launch video](https://www.youtube.com/watch?v=uZ72r5-B3UM) for a closer look.\n\nWondering has helped **more than 12k people complete over 42k lessons**, with 687 learning with us this past week and **100+ paying**. Start learning anything for free at[ https://wondering.app](https://wondering.app)\n\n**About the team**\n\nWe're Cheng-Wei (ex-NotebookLM) and Angelica (ex-Meta). We met at Cornell Tech over a shared passion for education and have been building together for 5+ years.\n\n**Our asks**\n\n* Give Wondering a try and let us know what you think[ https://wondering.app](https://wondering.app)\n* If there’s a group or organization you’d like to teach with Wondering, write to us at [founders@wondering.app](mailto:founders@wondering.app).\n\n![uploaded image](/media/?type=post&id=107951&key=user_uploads/1422401/d68e4385-976c-4eb4-b541-57e32716a14d)\n\n", "url": "https://www.ycombinator.com/launches/S59-wondering-duolingo-for-learning-anything", "votes": 154, "created_at": "2026-07-30T16:40:18.492Z" } ] }, { "id": 33319, "name": "Vernius Systems, Inc.", "slug": "vernius-systems-inc", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/b6b4e6b2f945df4b3400a22fc1d287a788bc4215.png", "website": "https://www.vernius.systems", "all_locations": "San Francisco, CA, USA", "long_description": "We make mass-deployable radars for unmanned systems. Our first product is Archimedes, a radar seeker that finds and locks onto targets, enabling autonomous intercept on any drone. We have over $580M in LOIs for Archimedes, across some of the largest defense primes in Ukraine, the EU, and the Middle East.\r\n\r\nSeekers are the most critical and costly component in every autonomous interceptor. Developing a radar seeker takes a 200+ person team at Raytheon and hundreds of millions of dollars. Archimedes is 100x cheaper than existing radar seekers on the market - and is the first product to make it feasible to deploy radars on every drone.", "one_liner": "Mass-deployable radars for air defense", "team_size": 3, "industry": "Industrials", "subindustry": "Industrials -> Defense", "launched_at": 1786812946, "tags": [ "Hard Tech", "Drones", "Radar", "Aerospace", "Defense" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Defense" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/vernius-systems-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/vernius-systems-inc.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/feed/update/urn:li:ugcPost:7497551262893383680/", "x_url": "https://x.com/Arthur_NC_/status/2091903458748375240", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/b6b4e6b2f945df4b3400a22fc1d287a788bc4215.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 366617, "full_name": "Arthur Nguyen-cao", "title": "Founder/CEO", "founder_bio": "Founder & CEO of Vernius Systems. 6+ years RF PCB engineering.\n\nLead 14 engineers at Defense Tech for Ukraine to develop FPV drone systems now fielded by Ukrainian units. Co-founded Auctify (raised $250K, featured in The Verge), where we built the world's first AI smart glasses. Developed the $2.6B USAF B-52 Commercial Engine Replacement Program proposal at Pratt & Whitney. Hobbies include computational fluid dynamics, rocketry, and civil aviation.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/arthur-nguyen-cao-690847146/", "x_url": "https://x.com/Arthur_NC_", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4373c5af412b516a25bd9d3afc5678c56a348e2e.jpg" }, { "user_id": 372571, "full_name": "Hisham El-Halabi", "title": "Founder/CTO", "founder_bio": "Founder at Vernius Systems. Previously founding engineer at Suave Money (HF0, Orange DAO-backed), building a non-custodial P2P fiat-to-crypto marketplace used by hundreds of thousands and adopted by YC startups Shor and Kontigo. Co-founder/CTO at Auctify, raising $77K in presales and pivoting to Sano AI calorie tracker (30K users, $2.5K MRR, $250K raised). BEng Engineering Science, University of Toronto.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/hisham-el-halabi-001/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/cd887780cb538a001e6e2bb72deb018117b59f82.jpg" }, { "user_id": 409168, "full_name": "Joe Gonzalez", "title": "Founder/COO", "founder_bio": "CoFounder at Vernius Systems. Serial founder: built Plucky Labs (backend-as-a-service for AI chat), XP Protocol (scaled to 200K+ users, 50K+ MAU), and Huddln (venture-backed Web3 startup). Previously Solutions Architect at Sandia National Labs under Q clearance, architecting secure comms for US Navy Tomahawk fleet. BS Computer Engineering, Sacramento State.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jge-2a7937115/", "x_url": "https://x.com/joegonzalez0886?s=11", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/677db7a1244f2604393fb1cc01f838806ea6304f.jpg" } ], "jobs": [ { "id": 111262, "title": "Radar PCB Engineer", "url": "https://www.ycombinator.com/companies/vernius-systems-inc/jobs/g2rbXLu-radar-pcb-engineer", "location": "San Francisco, CA, US / San Jose, CA, US", "type": "Full-time", "role": "Engineering", "role_type": "Electrical", "salary_range": "$150K - $350K", "equity_range": "1.00% - 3.00%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [ "PCB design", "Altium" ] } ], "news": [], "launches": [ { "id": 111298, "title": "Vernius Systems - Autonomous radar guidance for interceptor drones", "tagline": "Radar for interceptor drones", "body": "[**Vernius Systems**](https://www.linkedin.com/company/vernius-systems-inc/posts/?feedView=all)\n\n**TL:DR;** Vernius builds radar seekers that turn any drone into a fully autonomous interceptor. We build and test on the frontlines with operators, under real conditions. Because of this, we’ve secured **$500M** in LOIs for 2027 from top defense manufacturers and elite military units across the US, EU, and Ukraine.\\\n\\\n[https://drive.google.com/file/d/191oElqMvXaizOijw1ZN2aTY3ckYOGfVC](https://drive.google.com/file/d/191oElqMvXaizOijw1ZN2aTY3ckYOGfVC/view?usp=sharing)\n\n**Backed By:**\n\nOur early investors include the **former U.S. National Security Advisor** (LTG HR McMaster), the **Chief Engineer behind VBAT, Shield AI’s flagship product** (Austin Howard), and **world renowned Ukrainian Military analyst** (Rob Lee).\n\n**Problem:** \n\n99% of interceptors on the battlefield still require a pilot for manual guidance. Jet powered attack drones deployed by our adversaries have become too fast to intercept manually, and now account for over **50% of all attack drones**. Autonomous solutions, like the Patriot class missiles, cost $4M per shot.\n\n**Solution:**\n\nWe build a radar that turns any drone into a fully autonomous interceptor, reducing the cost per kill from **$4M to less than $4K**.\n\n**About us:** \n\nWe bring over a decade of RF PCB engineering experience. We've built critical software for the US Navy Tomahawk missile fleet, satellites in low Earth orbit, and FPV hardware fielded by frontline operators.\n\n![uploaded image](/media/?type=post&id=111298&key=user_uploads/409168/e627de41-d230-4320-80bb-90173e045879)\n\n**Left to right:** [Hisham El-Halabi](https://www.linkedin.com/in/hisham-el-halabi-001/), [Joe Gonzalez](https://www.linkedin.com/in/jge-2a7937115/), [Arthur Nguyen-cao](https://www.linkedin.com/in/arthur-nguyen-cao-690847146/)\n\n**Ask:**\n\nIf you’re developing interceptors (or you know someone developing them), we would love to chat.", "url": "https://www.ycombinator.com/launches/Sx8-vernius-systems-autonomous-radar-guidance-for-interceptor-drones", "votes": 10, "created_at": "2026-08-24T15:00:00.720Z" } ] }, { "id": 33320, "name": "OneCLI", "slug": "onecli", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/0405005492b7122ac5f4784e260fc64d7bd99053.png", "website": "https://onecli.sh", "all_locations": "San Francisco, CA, USA", "long_description": "OneCLI (YC S26) is an open-source platform for delegating business-critical work to AI agents, from Salesforce updates to data analysis and engineering investigations.\r\n\r\nAgents work across your company’s apps, prepare changes, and pause for approval when your rules require it. Your team controls what they can access and change, while credentials stay out of the agents’ reach and every call is logged.\r\n\r\nhttps://www.onecli.sh", "one_liner": "AI teammates for your business-critical work.", "team_size": 0, "industry": "B2B", "subindustry": "B2B -> Productivity", "launched_at": 1783888956, "tags": [ "AIOps", "B2B", "Workflow Automation", "Open Source", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Productivity" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/onecli", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/onecli.json", "year_founded": null, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/onecli", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/0405005492b7122ac5f4784e260fc64d7bd99053.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 406823, "full_name": "Jonathan Fishner", "title": "Founder/CEO", "founder_bio": "Co-founder of OneCLI", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jonathanfishner", "x_url": "https://x.com/jonathanfishner", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/6230ecfcb0be73a0475915f586f61d5139a12a58.jpg" }, { "user_id": 1935982, "full_name": "Guy Ben Aharon", "title": "Founder/CTO", "founder_bio": "Co-founder of OneCLI", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/guy-ben-aharon/", "x_url": "https://x.com/GuyBenAharon", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d2d3562621d02f714e16c1c63903eab6e5f71f23.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111185, "title": "OneCLI - Give every employee a secured, sandboxed pro assistant agent", "tagline": "Self-hosted agent harness for teams: sandboxed, policy-controlled, and never touching a real credential. Works from web and Slack.", "body": "Hi everyone 👋, we're Jonathan and Guy, founders of OneCLI.**\\\n\\\n🏆 TL;DR**\n\nOneCLI is an open source agent harness for teams. Each employee gets their own sandboxed pro assistant agent, connected to their GitHub, Gmail, Notion, Dropbox, or CRM right from the chat/Slack, while the whole workspace runs under one policy your admins control. \\\n\\\nAgents never hold real secrets: the credential is injected at the network layer, per request, after the call is authorized. You can't steal what isn't there.\n\n**Launch Video**\\\n\n\n🚧 **The Problem**\n\nLately, autonomous agents have arrived and shown up everywhere. Tools like OpenClaw and Hermes have become extremely popular and been adopted by open source communities, mostly by individuals. Companies, however, have been left behind: they need integration with their identity providers to enable easy provisioning of agents on behalf of an employee's identity, with least-privilege access.\n\nThose tools have fundamental flaws. Security fundamentals, like secret handling and organization-level control, are missing by default. So companies must choose between weak agents or making them powerful but risky.\n\n✅ **What OneCLI Does**\n\nOneCLI gives teams the whole harness, open source and self-hosted, so agents can be powerful and safe at once:\n\n* **Sandboxed agent per employee:** each agent runs isolated with its own memory, skills, and permissions.\n* **Secrets injected at the network layer:** the agent gets a placeholder, the gateway swaps in the real credential per request, after the call is authorized. Credentials never enter the agent's context, memory, or logs.\n* **Policy enforced outside the model:** admins define org-wide policy. Block endpoints and scope access per agent.\n* **Human-in-the-loop where it counts:** sensitive actions (send the email, delete the Linear ticket, empty the S3 bucket) wait for deterministic approval in the chat.\n* **Full identity trail:** every agent is bound to an employee, every call logged with who it acted for and which policy allowed it.\n\n🚀 **Team**\n\nWe've spent our careers in security. Jonathan built zero trust network access at Axis Security (acquired by HPE). The core ZTNA idea is that you never trust the client and enforce access outside it, at the network layer. Guy was the first employee at Argon (AppSec company acquired by Aqua Security), protecting client codebases. Agents need exactly the treatment ZTNA gives humans, so we built it.\n\n🙏 **Our Ask**\n\n* Do you want to provide each of your employees with a secure OpenClaw isolated in a sandbox? Shoot me an email: [jonathan@onecli.sh](mailto:jonathan@onecli.sh)\n* Star/try the repo:[ https://github.com/onecli/onecli](https://github.com/onecli/onecli). Self-host in minutes, free.", "url": "https://www.ycombinator.com/launches/SvJ-onecli-give-every-employee-a-secured-sandboxed-pro-assistant-agent", "votes": 29, "created_at": "2026-08-20T14:00:00.677Z" }, { "id": 106907, "title": "OneCLI - Your AI agents shouldn't be holding raw secrets", "tagline": "Agents hold placeholder tokens, real secrets get injected at the network layer", "body": "# **🏆 TL;DR**\n\n[**OneCLI**](http://onecli.sh) is an open-source credential gateway for **AI agents**.\\\nWhen an agent makes a call, OneCLI swaps in the real credential at the network layer and forwards it. The agent never touches the secret, **you can't steal what isn't there**.\\\n320K+ downloads, 2,500+ GitHub stars, and the default credential layer in NanoClaw (30K+ stars).\n\n**Ask:** If your team runs coding agents (Claude Code, Codex, Cursor) or autonomous agents in production. And if you can intro us to Head of DevOps or heads of DevSecOps at companies rolling out AI agents, email us at [jonathan@onecli.sh](mailto:jonathan@onecli.sh) \\\n\\\n**Sign up for free:** \n\n![uploaded image](/media/?type=post&id=106907&key=user_uploads/406823/5294ff50-4b88-4ee7-9c9c-3c776b5b45b5)\n\n\n\n# The Problem\n\nAI agents are doing real work now, which means they need real credentials: API keys, database passwords, OAuth tokens. And there's no safe way to hand them over.\n\nToday, every team does one of two things: hardcode secrets into **.env files** the agent can read, or build a one-off proxy hack. Both mean the **agent holds the secret**. An agent that holds a secret can leak it, through prompt injection, a malicious MCP server, a compromised dependency, or simply by pasting it into a log or a commit.\n\nSecrets management was designed for humans and **deterministic software**. Autonomous agents that read untrusted webpages, install packages, and write their own code are a fundamentally different threat model. The blast radius of one leaked production key is enormous, and teams are granting agents more access every week.\n\n# Our Solution\n\nOneCLI moves the credential out of the agent entirely.\n\nYou run one command, and every outbound call from your coding agent (Claude Code, Codex, Cursor, Cline) or autonomous agent (OpenClaw, Hermes, NanoClaw) routes through OneCLI. The agent only ever sees placeholder tokens. On each request, the proxy:\n\n1. Matches the request by host/path\n2. Checks it against policy\n3. Swaps the placeholder for the real credential\n4. Forwards the request\n\nThe agent never holds the secret, so there's nothing to steal, log, or exfiltrate. \\\nSecrets are AES-256-GCM encrypted at rest. The whole thing runs in a single Docker container with no external dependencies. It works with anything that can set HTTPS_PROXY.\n\nFor teams, we already support integration with IdPs and organization-wide policy setup, giving you full control and visibility over every access made by your agents.\n\n# Why this and not the PAM/identity vendors?\n\nBecause the credential problem for agents is a **network problem, not an identity problem**. Vault-based tools (1Password, HashiCorp Vault) still hand the secret to the agent to authenticate. The agent holds it, so the agent can leak it. OneCLI injects the credential at the wire: transparent injection, policy enforced on every outbound request, and the secret never enters the agent's context in the first place.\n\n# The Team\n\nWe've spent our careers in security. Guy (CTO) was the first hire at Argon (acquired by Aqua Security). Jonathan (CEO) worked on ZTNA and network security at Axis Security (acquired by HPE), after serving in Israeli Military Intelligence.\n\nWe hit this problem ourselves. Before OneCLI we built ChartDB together, an open-source database tool with 20K+ GitHub stars, and while building an agent orchestration layer on top of it, we kept hitting the same wall: there was no safe way to hand an autonomous agent credentials. Every team we talked to was either hardcoding API keys or duct-taping proxies. So we built the thing we wished existed.\n\n![uploaded image](/media/?type=post&id=106907&key=user_uploads/406823/ff451063-51d2-4c13-9c5f-cd1863b8c648)\n\n# The Ask\n\nIf you know a Head of DevOps or DevSecOps lead at a company rolling out AI agents, we'd love an intro.\n\n[jonathan@onecli.sh](mailto:jonathan@onecli.sh)", "url": "https://www.ycombinator.com/launches/RoJ-onecli-your-ai-agents-shouldn-t-be-holding-raw-secrets", "votes": 11, "created_at": "2026-07-23T13:44:46.846Z" } ] }, { "id": 33342, "name": "Nex", "slug": "nex", "former_names": [ "GarageSpace", "Inc.", "Nex", "Nex.ai" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/b8f563ce5c802b34a01e9ad0f5641eb4bb092c6c.png", "website": "https://nex.ai", "all_locations": "San Francisco, CA, USA", "long_description": "We build and deploy AI agents that manage prospecting, qualification, CRM hygiene, forecasting, lead routing, revenue recovery, and all the complex revenue work in between, on autopilot.", "one_liner": "Grok bot for GTM & Revenue teams", "team_size": 5, "industry": "B2B", "subindustry": "B2B -> Sales", "launched_at": 1784187262, "tags": [ "AIOps", "Workflow Automation", "Sales" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Sales" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/nex", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/nex.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/nexdotai/", "x_url": "https://x.com/nexdotai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/b8f563ce5c802b34a01e9ad0f5641eb4bb092c6c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 1371335, "full_name": "Najmuzzaman (Nazz) Mohammad", "title": "Founder/CEO", "founder_bio": "CEO/Co-founder at Nex (YC S26).\n\nFounding hire at VentureBasecamp, an ed-tech product acquired by Wadhwani Foundation within 2 years.\nEarlier founded a profitable e-commerce business during college.\nPreviously Senior PM at HubSpot and Clipboard Health.", "is_active": true, "linkedin_url": "https://linkedin.com/in/najmuzzaman", "x_url": "https://x.com/najmuzzaman", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5bd335c3878a06175e6944df3acdedf99db2ab1b.jpg" }, { "user_id": 3636126, "full_name": "Francisco Dias", "title": "Founder/CTO", "founder_bio": "Just trying to help out.\n\n4th engineer on HubSpot CRM (pre-IPO, ~0 customers)\nCo-wrote HubSpot's first component library\nFirst product-focused engineer on HubSpot's AI infrastructure team; helped grow it into a new group\nMentored 10+ ICs into technical leads\n\nHelped ship one of the most ambitious responsive websites, BostonGlobe.com\n\n22 marathons and a sub-11 Ironman in 22 months\n\nBOS > ATL > SFO", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/frandias/", "x_url": "https://x.com/ffrandias", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/fef4552be6d6bed0049c187c08b25e3fd19b792f.jpg" } ], "jobs": [], "news": [ { "title": "Show HN: A Karpathy-style LLM wiki your agents maintain (Markdown and Git) | Hacker News", "url": "https://news.ycombinator.com/item?id=47899844", "date": "Apr 24, 2026" } ], "launches": [ { "id": 111167, "title": "Nex - Claude Cowork for high-volume workflows, starting with GTM", "tagline": "AI that builds and runs the complex, high-volume GTM workflows only the world's best GTM engineers know how to build, at the price of a coffee.", "body": "**TL;DR:** Nex is Claude Cowork for high-volume workflows, starting with GTM. Describe a goal, and Nex builds an agent that runs it end to end across 1000+ integrations. The kind of work that would burn your entire weekly limit overnight on Claude or Codex runs on autopilot on Nex, for the price of a coffee, while that coffee brews. Try it at [nex.ai/signup](http://nex.ai/signup).\n\n _(Our AI-generated “The Office” parody video that went viral was taken down by YouTube. You can still [watch it here on LinkedIn](https://www.linkedin.com/posts/nexai-yc-s26-is-the-ai-gtm-engineer-that-ugcPost-7495214587882004480-Ch_b/?utm_source=social_share_send&utm_medium=member_desktop_web&rcm=ACoAAAseA8sB3xUI6osd_UWpcPBpx2tIFENh3N8)_)\n\n👋 Hey friends, we're Nazz and Francisco. We spent years at HubSpot running its three largest customer platforms, and we're backed by HubSpot founder Dharmesh Shah.\n\n**The problem: general-purpose agents aren't built for high-volume GTM**\n\nThe best GTM engineers can wire HubSpot, Clay, Outreach, Gong, and Slack into agents that run a revenue team. But run that same work through a general agent and it re-reasons every step, burns your entire weekly limit overnight, and still isn't reliable at volume. Almost nobody can hire enough GTM engineers to do it properly.\n\n**How Nex works**\n\nDescribe a goal, and Nex builds and runs the workflow across 1000+ integrations:\n\n* Lead generation, qualification, and enrichment\n* Scraping public records\n* Outbound over email and socials\n* CRM cleanup and hygiene\n* Re-engaging lost deals\n\nBecause most GTM work is repetitive and deterministic, Nex has AI write code for the predictable parts and run it fast without wasting tokens, saving AI judgment for the genuinely hard parts. That's how it breezes through high-volume work while not burning through your monthly AI budget in a day.\n\n**What that looks like**\n\nOn a single onboarding call, Nex audited thousands of records in a customer's HubSpot, flagged missing data, archived owners, bad associations, and zombie deals, and started fixing them before the call was over.\n\n**Why this matters**\n\nGTM is the beachhead. The same engine runs any high-volume workflow that's too slow and too expensive to run straight through a general model, and that pattern shows up in every operations-heavy function: finance reconciliations, support triage, recruiting outreach.\n\n**Our ask**\n\n* Try it free at [nex.ai/signup](http://nex.ai/signup)\n* If you run RevOps or GTM at a fast-moving company, we'd love to talk\n* If you know a RevOps or GTM leader, an intro would mean a lot: [founders@nex.ai](mailto:founders@nex.ai)", "url": "https://www.ycombinator.com/launches/Sv1-nex-claude-cowork-for-high-volume-workflows-starting-with-gtm", "votes": 7, "created_at": "2026-08-17T22:00:00.274Z" } ] }, { "id": 33344, "name": "Dawn Industries", "slug": "dawn-industries", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/be237324cf82c0125a34763f99ffb0cf35ca384d.png", "website": "https://www.dawnxr.com", "all_locations": "San Francisco, CA, USA", "long_description": "VIM connects to the entire cell: PLCs, robots, CNC machines, sensors. When a cell faults, VIM finds the cause, stages the fix, and the operator approves with one click. When a part drifts, VIM catches it seconds after the cut, inside the machine, and recommends the correction. Recovery drops from 30 minutes to under two, saving more than $15,000 per fault in large factories.", "one_liner": "Automatic Diagnosis & Fix for Industrial Automation Cells", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1786150873, "tags": [ "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/dawn-industries", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/dawn-industries.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/dawn-industries-inc/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/be237324cf82c0125a34763f99ffb0cf35ca384d.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 1720857, "full_name": "Joohyun (CJ) Cha", "title": "Founder/CEO", "founder_bio": "Founder at Dawn Industries (YC S26). Previously co-founder & CTO at Zipper, a creative collage social media backed by Unnoted Ventures (acquired by MetaZone). ML engineer at Avalo analyzing genomic data for crop development. Studied CS at Brown.", "is_active": true, "linkedin_url": "https://linkedin.com/in/joohyun-cha", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/02f28d9424990fbb7e48280e9e4384fa363a9c27.jpg" }, { "user_id": 2574843, "full_name": "Robin Lee", "title": "Founder/CTO", "founder_bio": "Founder at Dawn Industries. Previously founding engineer at NUUA, Aviation AI Startup that raised $10m.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/robinlee02/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e9ac55f690cab2a816d964d148195dd9d0f44a2e.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109744, "title": "Dawn Industries - Automatic Diagnosis & Fix for Industrial Automation", "tagline": "AI that understands the entire industrial cell and stages the fix in minutes.", "body": "Hey BF 👋 We are [@Joohyun (CJ) Cha](https://bookface.ycombinator.com/user/1720857) and [@Robin Lee](https://bookface.ycombinator.com/user/2574843), cofounders of [@Dawn Industries](https://bookface.ycombinator.com/company/33344). We’re building **VIM**, Automatic Diagnosis & Fix for Industrial Automation Cells.\n\n![uploaded image](/media/?type=post&id=109744&key=user_uploads/2574843/10398df4-83d1-4d24-860a-5ede83b91f9c)\n\n# **TL;DR:**\n\n**VIM is Sentry for industrial automation cells. When a cell faults, VIM reads the error code and the live machine signals, digs the root cause out of the OEM manuals and the cell's own history, and stages one safe fix for the operator to approve. When a CNC drifts, VIM catches it with the machine's own probe seconds after the cut and stages the offset correction. Downtime goes from hours to about a minute; scrap gets caught at part one instead of final inspection.**\n\nWatch how **VIM** works:\n\nRobot Cell Video: \n\nCNC Machine Video: \n\n# WHAT WE BUILT\n\nVIM sits on the cell and closes the loop from fault to recovery:\n\n1. **Connect**. Reads **PLC**, **HMI**, **robot**, **CNC**, **sensor**, and **driver state** without touching the controls you already trust.\n2. **Understand**. Matches the fault to live signals, the right page of the right manual, and this cell's own prior recoveries or trends the probe touches and traces drift to tool wear, thermal growth, a bumped fixture.\n3. **Recover**. Stages one bounded, validated action: a restart, a parameter change, an offset correction. The operator approves it like a pull request.\n\nBefore anything moves, the safety chain is verified: interlocks, light curtain, area scanner. VIM cannot bypass it; that's a precondition in the action policy, not a feature. And when the fault is genuinely hardware, VIM routes the incident to the right person with the exact diagnosis and repair steps attached, instead of a generic alarm and a prayer.\n\n**Autonomy is earned, not assumed**: VIM ships read-only, then one-click approval, then unattended recovery only on fault classes where it has earned trust. And it's mixed-vendor by design (FANUC robots and CNC controls, Universal Robots, Siemens, Allen-Bradley, Cognex, Keyence, Renishaw probes), because a cell is multi-vendor by definition and a copilot that speaks one vendor's language can't close the loop.\n\n![uploaded image](/media/?type=post&id=109744&key=user_uploads/2574843/83b531fe-5737-4f94-8b70-cd37330dae7c)\n\n# TEAM\n\n![uploaded image](/media/?type=post&id=109744&key=user_uploads/2574843/df76af6e-3516-4d76-9081-145424822ef3)\n\nWe're four people covering both halves of the problem.\n\n[Robin Lee](https://www.linkedin.com/in/robinlee02/) and [Joohyun (CJ) Cha](https://www.linkedin.com/in/joohyun-cha), cofounders, Robin was the founding AI engineer at an aviation AI startup, and built the production AI systems behind its $10M Series A and a 1st-place Minister's Award from Korea's Ministry of Science and ICT. CJ cofounded Zipper as CTO (acquired by MetaZone), did ML on genomic data at Avalo. We spent the past years in plants and machine shops learning all the faults and error that happen during manufacturing processes.\n\nOur founding engineers, come from Tesla and robotics research labs. They spent years building and fixing automation cells and various robot cases.\n\n# OUR ASK\n\n* **Intros to plant managers, manufacturing engineers, and controls engineers** at plants running robot cells (palletizing, case packing, welding, machine tending, inspection). If a stopped line ruins their week, we want to meet them.\n* **Intros to machining and quality leaders at plants and shops (primarily aerospace & defense) running CNCs**, anywhere scrap gets discovered at final inspection instead of at the machine.\n* **War stories.** Tell us about the cell that keeps stopping or the machine that keeps drifting. Best case, we make it your most reliable one.\n\nReach us at [joohyun_cha@dawnxr.com](mailto:joohyun_cha@dawnxr.com) or 401-226-3503, or book a demo at .", "url": "https://www.ycombinator.com/launches/SY4-dawn-industries-automatic-diagnosis-fix-for-industrial-automation", "votes": 12, "created_at": "2026-08-08T16:00:00.916Z" } ] }, { "id": 33345, "name": "CueBench", "slug": "cuebench", "former_names": [ "CueBench", "Snap Poker" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/8fc25d9ffc54b20b2ccee87e57733b12a6e8edf1.png", "website": "https://cuebench.dev", "all_locations": "San Francisco, CA, USA", "long_description": "", "one_liner": "RL post-training for science reasoning and performance engineering", "team_size": 3, "industry": "B2B", "subindustry": "B2B", "launched_at": 1786744876, "tags": [ "Reinforcement Learning", "Data Engineering" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/cuebench", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/cuebench.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/8fc25d9ffc54b20b2ccee87e57733b12a6e8edf1.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Vivian Midha Shen", "url": "https://www.ycombinator.com/people/vivian-midha%20shen" }, "founders": [ { "user_id": 3732725, "full_name": "Dillon Mehta", "title": "Founder/CEO", "founder_bio": "Founder and CEO at CueBench", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dillon-mehta-49aa2827b/", "x_url": "https://x.com/PickleDilll_", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/812fc022db435026c7bee56ba58f916bfd5218ee.jpg" }, { "user_id": 3762638, "full_name": "Neel Gadde", "title": "Founder/COO", "founder_bio": "Co-Founder and CRO at CueBench", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/neel-gadde-491880377/", "x_url": "https://x.com/neelapolitan?s=11", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d2be7502db72b4faa82985dd5181743eabe8a5aa.jpg" }, { "user_id": 3791688, "full_name": "Rishan Hemrajani", "title": "Founder", "founder_bio": "Co-Founder & CTO of CueBench.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rishan-hemrajani-254bba292", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e3bd268e04f9551b93d77346c5652657d25a2b69.jpg" } ], "jobs": [], "news": [], "launches": [] }, { "id": 33348, "name": "Autoresearch, Inc.", "slug": "autoresearch-inc", "former_names": [ "hiloop" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/5b91efea37962386a163af58e2f2920b1e9d0aa7.png", "website": "https://autoresearch.inc", "all_locations": "San Francisco, CA, USA", "long_description": "Autoresearch helps teams train agents for tasks where general models are not good enough. Give us a task, your current agent or model, and an evaluation. Autoresearch runs an autoresearch campaign across data, SFT and other post-training methods, continual learning, prompts, tools, harnesses, and systems, then returns the best verified improvement. It runs hosted or in your cloud.\r\n\r\nWe provide the research system around models: persistent memory, full experiment lineage, compute orchestration, and statistical verification. We’re starting with agent and model training, continual learning, and optimization.\r\n\r\nReach out to us for early access at founders@autoresearch.inc.", "one_liner": "Autoresearch as a service: we solve your hardest measurable problems", "team_size": 3, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1785348094, "tags": [ "Developer Tools", "Machine Learning", "Reinforcement Learning", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/autoresearch-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/autoresearch-inc.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/autoresearchinc/", "x_url": "https://x.com/hiloopai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/5b91efea37962386a163af58e2f2920b1e9d0aa7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 1714115, "full_name": "Karan Brar", "title": "Founder/CEO", "founder_bio": "Founder at Autoresearch, Inc. Previously ML at Reducto, Head of ML Infra at DynamoAI.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/brarkaran", "x_url": "https://x.com/deepmatmul", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e596ed985a539262821de87d80ea418725ab7a90.jpg" }, { "user_id": 1597736, "full_name": "Thomas Boser", "title": "Founder/CTO", "founder_bio": "Founder at Autoresearch, Inc. Previously engineer at Reducto, founding engineer at Crosswise, senior ML engineer at Discord, ML at Sentropy & SoFi.", "is_active": true, "linkedin_url": "https://linkedin.com/in/tboser", "x_url": "https://x.com/thomasboser", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/7c1d2d3062a34b6d55d2d4e7aef55e887853400b.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108337, "title": "hiloop: we run thousands of experiments to improve your models", "tagline": "Send us a hard task, your model or agent, and an eval. We run the research campaign and return a verified improvement.", "body": "We're Karan and Thomas, and we're building hiloop: **infrastructure for automated research**.\n\n**TL;DR: we gave two stock coding agents 50 B200s and our infra. They ran 4,188 experiments in two days and beat the published state of the art on [Karpathy's autoresearch benchmark](https://github.com/karpathy/autoresearch). The same machinery found a [NanoGPT speedrun](https://app.primeintellect.ai/speedrun/nanogpt) candidate that beats the best known result. [See](https://hiloop.ai/blog/search-is-enough/) full writeup.**\n\nNow we want to point it at your hardest problems.\n\n![uploaded image](/media/?type=post&id=108337&key=user_uploads/1714115/63396f7f-3e21-41ea-83ea-54cbaaa9f63f)\n\n**The results**\n\nOn Karpathy’s autoresearch benchmark, our final recipe reached **0.9016 val_bpb**. The previous best published result was [Recursive’s 0.9109](https://www.linkedin.com/pulse/first-steps-toward-automated-ai-research-recursive-si-kybtc) on B200 hardware. For NanoGPT, our implementation had **12.54% lower runtime** than upstream.\n\nThere was no bespoke research agent or elaborate scaffold. We used Claude and GPT-5.5 out of the box. What changed was the system around them.\n\n**What hiloop does**\n\nGive us a hard task, your current agent or model, and an evaluation criteria. Hiloop runs an autoresearch campaign across training, data, prompts, tools, harnesses, and systems, then returns the best verified improvement. It runs in your cloud or hosted by us.\n\nWe’re starting with agent and model training: SFT, post-training, continual learning, and optimization.\n\n[https://drive.google.com/file/d/1Rv8HttSHZ-yvSJk6hsWQsqCXWTHJQogK](https://drive.google.com/file/d/1Rv8HttSHZ-yvSJk6hsWQsqCXWTHJQogK/view?usp=sharing)\n\n**Who we are**\n\nWe met at Reducto, where we built its ML, post-training, and platform systems. Karan previously led ML infrastructure at DynamoAI . Thomas was a founding engineer at Crosswise and an MLE at Discord and SoFi.\n\n**Two asks**\n\n1. **Want your models or agents better, faster, or cheaper?** If you have a hard objective in post-training, continual learning, inference optimization, or classic ML, send us the task, your baseline, and the eval. We’ll personally run a campaign with you through verified improvement.\n2. **Building autoresearch or RSI loops?** Reach out to give our system a try! We’d love to compare notes.\n\nEmail [**founders@hiloop.ai**](mailto:founders@hiloop.ai) to get in touch with us!", "url": "https://www.ycombinator.com/launches/SBN-hiloop-we-run-thousands-of-experiments-to-improve-your-models", "votes": 8, "created_at": "2026-07-29T19:50:09.250Z" } ] }, { "id": 33349, "name": "Robocurve", "slug": "robocurve", "former_names": [ "Telos", "Robocurve PBC" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/1cfdafb2f71eef92d986496e842d5218b330742b.png", "website": "https://robocurve.org", "all_locations": "San Francisco, CA, USA", "long_description": "Robocurve builds open-source tools and independent benchmarks to measure how well robots can do real-world jobs. Instead of relying on unverified demo videos from frontier labs, we score their models on reproducible benchmarks that anyone can trust.\r\n\r\nToday there are no well-run, standardized robotics benchmarks. Labs evaluate in-house, and no independent group has stepped in to run continuous benchmarking as a service. The result is that no one actually knows how good anyone else is, or where the real frontier sits.\r\n\r\nGood benchmarks require operating and maintaining physical hardware and real-world setups. Simulation only goes so far, since models that look strong in sim can show large performance gaps once deployed in the real world. Building real-world benchmarks means coordinating job-domain experts, evals engineering, and hands-on robotics all at once.\r\n\r\nFrontier labs are targeting general-purpose robotics by 2028, yet the field of robotics evals barely exists. Whoever builds the trusted measure of robot capability becomes the reference everyone relies on. We combine backgrounds in AI evals and robotics to build and grow this field as fast as possible. We already shipped v1 of our open-source framework, Inspect Robots, and ran our first pilot scoring a frontier model on a real robot.", "one_liner": "Evals for robots", "team_size": 3, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1783378088, "tags": [ "Hard Tech", "Hardware", "Robotics", "Open Source", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/robocurve", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/robocurve.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/robocurve/", "x_url": "https://x.com/robocurve", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/1cfdafb2f71eef92d986496e842d5218b330742b.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Ankit Gupta", "url": "https://www.ycombinator.com/people/ankit-gupta" }, "founders": [ { "user_id": 2392672, "full_name": "Jay Chooi", "title": "Founder/CEO", "founder_bio": "Founder and CEO of Robocurve. MA Statistics and BA CS/Math from Harvard. Jay builds evals of frontier AI and forecasts its progress. Previously Research Fellow at MATS, researcher at the UK AI Security Institute, and top contributor to Inspect Evals, the UK government's AI eval framework. Published at ACM EC, ICML, ACL, and EMNLP. Called the 2024 election correctly in all 50 states. Won a Rhodes Scholarship. Gold medal at the International Olympiad on Astronomy and Astrophysics.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/jeqcho", "x_url": "https://x.com/chooi_jeq", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/5df96d3283ec234b3a09c35933b2582a7ef1a067.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 112513, "title": "Robocurve — Real-World Evaluations of Physical AI", "tagline": "Real-world evals for robots", "body": "Hi, I’m [Jay](https://www.linkedin.com/in/jeqcho/), co-founder of [Robocurve](http://robocurve.org). We measure and reports what frontier robots can actually do in the real world.\n\nOur launch video \n\nFrontier labs are racing to build general-purpose robots within two years, yet no one knows what today's robots are truly capable of. Demo videos of robots performing narrow, cherry-picked tasks proliferate, but there are no rigorous, continuous, standardized benchmarks for robotics.\n\nRobocurve builds benchmarks for robots and the AI models that control them, to help the world understand how far robotics has come and how fast it's moving. Our benchmarks range from playing Jenga and making a sandwich to constructing a data center.\n\nGeneral-purpose robots could arrive before the end of the decade, with far-reaching implications: accelerated data center buildouts, explosive economic growth from a fully automated economy, and intensified geopolitical competition between the US and China. Real-world evaluations from Robocurve will help clarify the timeline to general-purpose robots so that society can prepare for their arrival well ahead of time.\n\n[Jay](https://www.linkedin.com/in/jeqcho/) is a Rhodes Scholar and a top contributor to UK AI Security Institute's Inspect Evals. He has published papers in ICML, EMNLP, ACL, ACM EC and his research has been featured in MIT Technology Review. Previously, he spent time at AWS Bedrock, MATS, and Harvard where he completed his BA and MA with a highest honors thesis on AI and democracy.\n\nIf you build robots or AI models, reach out! We'd love to put them to the test. Learn more at [robocurve.org](http://robocurve.org).", "url": "https://www.ycombinator.com/launches/TGj-robocurve-real-world-evaluations-of-physical-ai", "votes": 12, "created_at": "2026-08-27T20:30:00.743Z" } ] }, { "id": 33408, "name": "Salem Robotics Inc", "slug": "salem-robotics-inc", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/fbf386dc8fa91340ed1cce1b80613f66e88ee12d.png", "website": "https://salemroboticsinc.com", "all_locations": "San Francisco, CA, USA", "long_description": "Our software turns any robot we deploy into an everyday independent worker at your facility. They are able to walk sites, handle instruments, overcome obstacles, collect measurements, and generate auditable compliance records, replacing hours of mandatory manual surveys and inspections. Built by Los Alamos robotics veterans and deterministic-manipulation researchers, we started in nuclear and are already expanding to other hazardous industries.", "one_liner": "Deploying robots to replace workers in hazardous spaces", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Manufacturing and Robotics", "launched_at": 1783052221, "tags": [ "Hard Tech", "Robotics", "Energy", "Defense", "Industrial" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Manufacturing and Robotics" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/salem-robotics-inc", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/salem-robotics-inc.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/salem-robotics/", "x_url": "https://x.com/salemroboticshq", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/fbf386dc8fa91340ed1cce1b80613f66e88ee12d.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/4593ea24a36e629876ca1a48e9e7db1a683530c2.png" ], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 3403013, "full_name": "Caleb Horan", "title": "Founder/CEO", "founder_bio": "Co-founder and CEO at Salem Robotics. Previously a published robotics PhD student at UT Austin, developing task planning for autonomous systems. NRC Fellow. Deployed multi-robot teams at Los Alamos National Laboratory's Plutonium Division. MS in Robotics from UT Austin.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/caleb-horan", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e7cff240bcb1484bcdff91bb1419d9c25f3162cf.jpg" }, { "user_id": 3208433, "full_name": "Janak (Crasun) Panthi", "title": "Founder/CTO", "founder_bio": "Co-founder and CTO at Salem Robotics Inc. PhD in Robotics from UT Austin (Dec. 2026). 6 yrs industry experience leading product development and deploying quadruped robots in nuclear facilities. US Army.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/janak-panthi", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/19b6a7b10bbd6739abbb938281eb9c068eebf9c0.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108976, "title": "Salem Robotics - Autonomous robots for hazardous spaces", "tagline": "We deploy robots to perform autonomous inspections and surveys in hazardous spaces, like nuclear. ", "body": "We're [Caleb](https://www.linkedin.com/in/caleb-horan) and [Janak](https://www.linkedin.com/in/janak-panthi), PhD roboticists from UT Austin. We've spent a combined 15 years in nuclear, 10 of which deploying autonomous robots at Los Alamos National Lab, and we're building Salem Robotics.\\\n**TLDR:** Our software turns robots into independent survey and inspection workers at your facility.\n\n[Launch Video](https://youtu.be/7xObOE6knZM)\n\n## **Life and death pain worth $Millions/year**\n\nAt every nuclear facility, highly trained radiation-protection teams spend hours each day walking the site, filling out hand-drawn maps, and searching for something life-threatening: radiation. IN THE BEST CASE they find nothing, and still have to return to their posts to transcribe the data into a computer before doing the actual high-value work they were assigned. These surveys are time-consuming, dangerous, expensive, and prone to data inconsistencies that carry direct regulatory consequences.\n\nThe same pain exists well beyond nuclear. Whether it's gas-leak inspection or ammonia surveys, similar workflows run across oil, gas, chemical, biological, space, and other hazardous facilities.\n\n## **What we do**\n\nAn operator defines the work from a simple dashboard; the robot then navigates the site, handles instruments, opens doors, turns valves, collects measurements, and generates structured, auditable, compliance-ready reports. We're hardware-agnostic, so plants run our software on robots they already own, many of which sit in a closet today as \"$200k paperweights, \" or we can provide the right robot for the job. Our stack is designed for safety-critical infrastructure and comes with theoretical safety guarantees, exactly what regulators and plant leadership require. Our core manipulation framework was not possible before and is based on **breakthrough founder research** published and presented at the most prestigious IEEE journals and conferences in the world. \n\nWe're already in active contract-scoping conversations across the nuclear industry and oil & gas.\n\n**Our asks**\n\n* Intros to operations or innovation leaders in nuclear or similar hazardous sites (oil, gas, chem, bio, space).\n* Advice from founders who've sold into safety-critical industrial environments.\n* Sites where routine surveys are frequent, hard to staff, or particularly painful? We'd love to hear about them.\n\nComments, Emails, and DMs welcome. \\\n[Engage@salemroboticsinc.com](mailto:Engage@salemroboticsinc.com)", "url": "https://www.ycombinator.com/launches/SLg-salem-robotics-autonomous-robots-for-hazardous-spaces", "votes": 29, "created_at": "2026-08-04T19:00:03.963Z" } ] }, { "id": 33441, "name": "Allia Health", "slug": "allia-health", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/3d6f40f216c0da6b1f8b4caa594f59a5b194dd82.png", "website": "https://allia.health/", "all_locations": "San Francisco, CA, USA", "long_description": "Allia is the first full-stack, AI-native clinical group for mental health. We bring independent practices into one national clinical group that can deliver, measure, and improve care at scale .", "one_liner": "AI-Native Medical Group for Mental Health", "team_size": 3, "industry": "Healthcare", "subindustry": "Healthcare -> Healthcare Services", "launched_at": 1787168275, "tags": [ "Healthcare", "Mental Health", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Healthcare", "Healthcare Services" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/allia-health", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/allia-health.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/alliahealth/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/3d6f40f216c0da6b1f8b4caa594f59a5b194dd82.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 3759395, "full_name": "Amie Leighton", "title": "Founder/CEO", "founder_bio": "Co-founder & CEO of Allia Health, the first AI-native medical group for mental health. Oxford dropout (declined med school); ex-Head of Healthcare M&A, Oxford M&A Society.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/amieleighton/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c1f94471f56e5b0cd389c75f662b942825fde549.jpg" }, { "user_id": 3759416, "full_name": "Saroosh Khan", "title": "Founder/CTO", "founder_bio": "Scored 94.2% on ARC-AGI-2, the leading AI reasoning benchmark (first to cross the 85% \"human-like reasoning\" threshold, next best was 74% by DeepMind) \n\nOne of the 3 individuals to receive a national scholarship\n", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/saroosh-khan/", "x_url": "https://x.com/sarooshkhan98", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/33b5c3d0e55f2dbc77b70ec81e3a8f43de841cdd.jpg" } ], "jobs": [ { "id": 112026, "title": "Growth / Marketing Lead", "url": "https://www.ycombinator.com/companies/allia-health/jobs/UDZwxH6-growth-marketing-lead", "location": "San Francisco, CA, US", "type": "Full-time", "role": "Marketing", "role_type": null, "salary_range": "$90K - $140K", "equity_range": "", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 111428, "title": "Allia: The first full-stack, AI-native mental health group", "tagline": "8000+ independent mental health clinicians in America already run their practice on Allia. We turn that network into one national clinical group that delivers coordinated, measurable care and contracts with the country's largest payers.", "body": "**TL;DR:** Allia is the first full-stack, AI-native mental health group. We have **$50M+ in annualized national insurance contracts** and will begin delivering care in September 2026.\n\nWe bring independent practices into one national clinical group that can deliver, measure, and improve care at scale: **600+ clinicians across 70 locations in 32 states, plus telehealth nationwide.**\n\nEvery clinician runs on a single intelligent system of record powered by collective intelligence at a scale, no independent practice could achieve alone.\n\n\n\n### **How we got here**\n\nWe have the fastest-growing distribution base in behavioral health: 8000+ independent mental health clinicians in America already run their practice on Allia. That's 1,000+ practices every day, caring for 300,000+ patients. It's our wedge: clinicians adopt the OS to run their practice, and we convert those practices into our clinical group.\n\n### **What Allia does**\n\nAllia is a clinical group that gives independent mental health practices the coordination, infrastructure, and payer contracts of a connected health system,  while letting them stay independent. \n\n1. **One coordinated group.** Because every practice operates as part of Allia, care no longer stops at the walls of a single office. Patients move through therapy, psychiatry, primary care, specialty, and higher levels of care as one coordinated journey across a single entity.\n2. **Care that stays connected.** Allia matches each patient to the right clinician and coordinates what comes next across the full continuum of care. Referrals are tracked through completion, so patients are never left to navigate the system alone.\n3. **Operations that run in the background.** Eligibility, intake, scheduling, documentation, coding, billing, follow-ups, and payer reporting happen automatically as care is delivered. Far less gets chased by hand, so clinicians spend their time on patients, not paperwork.\n4. **Outcomes that become visible.** Measurement-informed care, powered by our proprietary agentic technology, is built into every patient's journey, identifying who is improving, who is disengaging, and who needs a change in treatment. The group turns that evidence into better care across the continuum, and into the outcome data that earns better payer contracts.\n\n### **Traction**\n\n* $50M+ in annualized care capacity under signed national contracts with America's largest health insurance companies\n* 8,000+ clinicians across 1,000+ practices use Allia every day (our distribution wedge); 600+ are already activated into the clinical group.\n* 300,000+ patients supported across the network\n\n### **Team**\n\n* **Amie Leighton (CEO):** dropped out of Oxford and declined a place at medical school to build Allia.\n* **Saroosh Khan (CTO):** 94.2% on ARC-AGI-2 (next best: 74%, Google DeepMind); first to cross the 85% \"human-like reasoning\" threshold. Built Allia's deep-reasoning population-health agent, the first inside a system of record for medicine.\n* **Clinical team:** former CMO of the nation's largest behavioral health company **(IPO >$8B)**, former CMO of the country's **3rd-largest health plan,** and leaders who scaled the biggest names in behavioral health.\n\n### **What we're looking for**\n\nIntros to health systems and mental health practices who want to join with the group. Reach out: [team@allia.health](mailto:team@allia.health)", "url": "https://www.ycombinator.com/launches/SzE-allia-the-first-full-stack-ai-native-mental-health-group", "votes": 162, "created_at": "2026-08-20T12:30:04.031Z" } ] }, { "id": 33540, "name": "Osseus", "slug": "osseus", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/fd39ca163a2a8ea396a6ec66701926317f9f416c.png", "website": "https://osseus.ai/", "all_locations": "London, England, United Kingdom", "long_description": "In an era of cheap intelligence, the most important application of AI will be to improving human health, globally. Osseus is building the frontier medical AI lab, starting with the clinical datasets, evaluations, and reinforcement learning environments that help AI labs train more capable, aligned, and safe models.", "one_liner": "The frontier data platform for aligned health and biological AI", "team_size": 2, "industry": "B2B", "subindustry": "B2B", "launched_at": 1783740069, "tags": [ "B2B", "Biotech", "Healthcare", "Data Engineering", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United Kingdom", "Europe" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/osseus", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/osseus.json", "year_founded": 2026, "city": "London", "country": "GB", "linkedin_url": "https://www.linkedin.com/company/osseus-intelligence", "x_url": "https://x.com/osseusinc", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/fd39ca163a2a8ea396a6ec66701926317f9f416c.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 1409413, "full_name": "Anish Mariathasan", "title": "Founder/CEO", "founder_bio": "Co-Founder at Osseus (YC S26). Prev. co-founded a compliance AI startup. ML research at a Harvard-HKU joint lab on computer vision for cancer research. MEng in Biomedical Engineering from Imperial College London, dropped Oxford PhD in AI for Health. Awarded academic scholarships for both.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/anishmariathasan/", "x_url": "https://x.com/mariathasanish", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/2e6cc631f0d2a1f1b51513c008b98bed931bad1b.jpg" }, { "user_id": 1979222, "full_name": "Ayush Patel", "title": "Founder/CTO", "founder_bio": "Co-Founder at Osseus (YC S26). 2x intern at Apple, where he built an agentic QA system. MEng in Computer Science from Imperial College London.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ayush-patel4/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ccdce4615ca525f275d384024cd3e6eb8abc2e18.jpg" } ], "jobs": [], "news": [], "launches": [] }, { "id": 33573, "name": "hardware intelligence", "slug": "hardware-intelligence", "former_names": [ "Viki", "hi", "Hardware Intelligence" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/3c2193e7412febe2e812491e1a73617bb09eb897.png", "website": "https://hardwareintelligence.ai", "all_locations": "San Francisco, CA, USA", "long_description": "Hardware Intelligence makes chip design tools that cut simulation and debugging time by an order of magnitude. When a chip test fails, engineers need a recording of every internal signal - a waveform. We’ve developed WaveZip, which captures waveforms 500x smaller with 200x less overhead, and Wave, which beats frontier models on industry chip debugging benchmarks. We’re Rishov Sarkar (Georgia Tech Ph.D. in chip simulation) and Athreya Anand (ex-Google, M.S. ML/AI), YC S26. https://hardwareintelligence.ai", "one_liner": "AI tools for the chips your AI runs on", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1784052551, "tags": [ "Developer Tools", "Hardware", "Enterprise Software", "Semiconductors", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/hardware-intelligence", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/hardware-intelligence.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/hardwareintelligence/", "x_url": "https://x.com/hwintelligence", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/3c2193e7412febe2e812491e1a73617bb09eb897.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/84867b76e94b5d17051762179fca04132eaaa7be.png", "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/c8cf823528ccb123bf55a70cc8e1a177803988de.png" ], "primary_partner": { "name": "Jon Xu", "url": "https://www.ycombinator.com/people/jon-xu" }, "founders": [ { "user_id": 551607, "full_name": "Athreya Anand", "title": "Founder/CEO", "founder_bio": "Co-founder & CEO of Hardware Intelligence. Athreya was a tech lead at Google where he led AI efforts across orgs from agentic workflow management tools to prediction models for Google Ventures. He has a Master's from Georgia Tech in Machine Learning and A.I., with technical expertise in agentic tools and chip design through his time at Google, AWS Inferentia, and Tesla. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/athreyaanand", "x_url": "https://x.com/athreyaanand", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d015fb0c86a6bdae44f5b5c36e2b3d18f086454a.jpg" }, { "user_id": 2565323, "full_name": "Rishov Sarkar", "title": "Founder/CTO", "founder_bio": "Co-founder & CTO at Hardware Intelligence. Rishov graduated with his PhD in Electrical & Computer Engineering from Georgia Tech with six published first-author/co-first-author conference papers, including one best paper runner-up. He also developed simulators for AMD's AI Engine accelerator and worked on LLVM-based compilation flows at Siemens EDA. He earned his MS and BS in Computer Engineering from Georgia Tech.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/rishovsarkar/", "x_url": "https://twitter.com/ArkaneMoose", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/d7666af58a4a49fbf9b8355cdffcb981f5df6505.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107661, "title": "hardware intelligence: AI tools for the chips your AI runs on", "tagline": "Bringing the $20B chip tooling industry into this decade", "body": "---\n\n---\n\n**Update:** We’ve just launched our second tool, WaveZip - a genuine breakthrough method for capturing and storing waveforms. Read more at \n\n---\n\n---\n\n\n\n—\n\nWe’re Athreya and Rishov - the founders of [@hardware intelligence](https://bookface.ycombinator.com/company/33573). We’re here to pick a fight with the $20B industry whose debuggers suck and won’t open until an ancient license server gives it permission.\n\n**TL;DR: Our first product, Wave, is a SOTA waveform debugger that lives completely in the terminal, built for the agentic era.** An agent can jump you to the moment a design breaks, drag in waveforms and manipulate them, trace bad signals to its source and even prove its fix against the simulation - all visual. We allow teams to build better chips in less time.\n\n**The quick ask:** If you or anyone you know builds or verifies chips (ASICs/FPGAs), please email us! We would love to talk :)\n\n\n\n❌ **The Problem**\n\nThe AI boom is a chip boom - more companies and teams than ever are designing silicon. However, the reality is hardware is not software: you can’t patch a chip after it ships. A bug that reaches manufacturing can be a nine-figure mistake. That’s why **companies spend 70% of the entire chip design process just on verification**.\n\nHardware verification engineers run thousands of simulated tests first on their code (simulation) which typically can take **hours to days**. If one of these tests fail, debugging means reading a waveform - a log of every signal, millions of ticks long, hiding the tick where a value went wrong. It’s literally a needle in a haystack… a big haystack. It’s common for these waveform files to be several gigabytes large. **Finding this needle is the single largest activity in verification.**\n\n**The current tools are dinosaurs.** 20-year-old GUIs behind license servers whose seats run six figures a year sold by our favorite incumbents: Synopsys, Cadence, and Siemens. Free tools are even worse.\n\n![uploaded image](/media/?type=post&id=107661&key=user_uploads/551607/fd100b13-f0dd-45f0-8ee4-395edf96fcae)\n\nYes. This is from 2026.\n\nSoftware engineers got coding agents. Chip engineers got a queue for a license and whatever this is.\n\n👋 **What Wave actually does**\n\n![uploaded image](/media/?type=post&id=107661&key=user_uploads/551607/a75818b0-b001-466b-9586-705d26deb5f9)\n\nWave is the first of its kind! **A fully terminal-native waveform debugger with a verification agent built in.** Wave opens files before Verdi, the industry-standard viewer, finishes its license check. Let’s do a quick dive!\n\n![uploaded image](/media/?type=post&id=107661&key=user_uploads/551607/ac7e5c75-8abb-4a6d-881f-a2a9b883a8b3)\n\n**Why the terminal?** Surprisingly, most hardware engineers are terminal power users - typically doing all their coding in Vim. We wanted to meet them where they are and make our product extremely intuitive to use - we even employ Vim bindings across the interface.\n\n![uploaded image](/media/?type=post&id=107661&key=user_uploads/551607/d4063fbc-2f57-4b5c-a421-bfaa697433d6)\n\n**Ask in plain english.** _Why is a wave x at 12.4us?_ The agent will pull in signals, act on them and trace the root cause. Everything happens in front of your eyes.\n\n![uploaded image](/media/?type=post&id=107661&key=user_uploads/551607/8f37dd5c-22f0-4271-b665-a096c5dbfbd2)\n\n**Pick your fix and see it.** Wave will propose different approaches with tradeoffs. The definition of “best” in hardware design depends on the use case. Maybe you want to optimize for performance? Maybe for power? Pick the one for you! Wave re-simulates and re-reads the new waveform. See a diff of how the waveform changed and a visual indication of where the fix was implemented.\n\n![uploaded image](/media/?type=post&id=107661&key=user_uploads/551607/b0875bc9-fb55-4275-8c7e-11f926bc597d)\n\nNow watch it fix a bug end to end with Rishov!\n\n\n\n🙏 **Asks & Offers**\n\nIf you design chips or lead a team, email us! We’d love to work with you and speed up your RTL development.\n\nKnow someone who does? Email us!\n\nJust want to chat about hardware? **EMAIL US!!**\n\n👬 **The Team**\n\nWe met in our freshman dorm at Georgia Tech nearly a decade ago! We became best friends and roommates for all four years of college - after, Athreya went to NYC to work at Google and Rishov stayed to do his PhD. We promised each other that once we’re free, that we’ll go all-in and work on something we’re passionate about together. Turns out, Rishov’s deep PhD thesis on chip simulation and compilers was the perfect spark.\n\nToday we announce our company and are roommates once again ❤️\n\n![uploaded image](/media/?type=post&id=107661&key=user_uploads/551607/eef30a82-c606-494a-a093-3649920c7857)\n\n—\n\nSay hi 👋 [hi@hardwareintelligence.ai](mailto:hi@hardwareintelligence.ai) · [hardwareintelligence.ai](http://hardwareintelligence.ai)\n\n[@Athreya Anand](https://bookface.ycombinator.com/user/551607) · [athreya@hardwareintelligence.ai](mailto:athreya@hardwareintelligence.ai)\n\n[@Rishov Sarkar](https://bookface.ycombinator.com/user/2565323) · [rishov@hardwareintelligence.ai](mailto:rishov@hardwareintelligence.ai)\n\nP.S. Wave is just the start 👀 We got some super cool things coming to close the loop of hardware engineering. ", "url": "https://www.ycombinator.com/launches/S0T-hardware-intelligence-ai-tools-for-the-chips-your-ai-runs-on", "votes": 50, "created_at": "2026-07-27T16:00:24.675Z" } ] }, { "id": 33606, "name": "Billow AI Labs", "slug": "billow-ai-labs", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/2c42b391e7d2cf9beb71d20f5ce1ed2b4e77e77d.png", "website": "https://thebillow.ai/", "all_locations": "San Francisco, CA, USA", "long_description": "Billow handles the work of Accounting and FP&A FTEs keeping teams compliant and audit-ready. We work within existing stacks (NetSuite, Coupa, and 80+ other systems) and complex spreadsheets, chasing down information and coordinating approvals across stakeholders. \r\n\r\nBuilt by repeat founders; Joan built and exited an accounting/ERP platform (45,000 businesses in 4 years) and Phil built Bloomberg's financial “brain” data infrastructure, and as CTO of a trading platform ($1B in volume). Advisors include CFOs, controllers, auditors from publicly-traded biopharma, and a Big-4 Accounting firm. ", "one_liner": "AI-native Accounting Firm that's killing Deloitte", "team_size": 0, "industry": "B2B", "subindustry": "B2B -> Finance and Accounting", "launched_at": 1784327267, "tags": [ "Finance", "B2B", "Workflow Automation", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Finance and Accounting" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/billow-ai-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/billow-ai-labs.json", "year_founded": null, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/billow-ai/", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/2c42b391e7d2cf9beb71d20f5ce1ed2b4e77e77d.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/ce73d929ec37820269e48e542e1af8c6deaf9d43.png" ], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 983067, "full_name": "Joanathan McIntosh", "title": "Founder/CEO", "founder_bio": "Co-founder & CEO of Billow AI Labs. Repeat, exited founder - built an accounting/ERP software (45,000 businesses in 4 years). Formerly en-route to become a CPA/CFA, ex-Merrill Lynch and StreetLight Data (AI/ML for the U.S. government and regulated industries).", "is_active": true, "linkedin_url": "https://linkedin.com/in/joanathanlim", "x_url": "http://x.com/joanathanlmc", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/c7bb76a8542e736d8c8e672a059c389f9decd761.jpg" }, { "user_id": 548009, "full_name": "Philip Moniaga", "title": "Founder/CTO", "founder_bio": "Co-founder & CTO of Billow AI Labs. Phil (YC W22) had built Bloomberg’s finance \"brain\" data infra and co-founded/CTO of a trading platform ($1B+ in volume)", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/philipmoniaga/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/0748bfa571fe070604aad2bc278c30312f1a1d5b.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110394, "title": "Billow - Killing Deloitte with AI", "tagline": "AI-native Accounting Firm for Enterprise", "body": "# **_We’re killing Deloitte with AI_**\n\n[https://drive.google.com/file/d/1YaUyJE-ZL0poxDAFGv962RniFsue5nyE](https://drive.google.com/file/d/1YaUyJE-ZL0poxDAFGv962RniFsue5nyE/view?usp=sharing)\n\n**Accounting is a $700B industry that hasn’t changed in 200 years.** Today, 80% of companies still do their accounting by hand – manually exporting data, building spreadsheets, and chasing approvals.\n\n**Billow is an accounting firm that replaces human accountants with AI agents.**\n\nWe connect to your existing tools (NetSuite, QuickBooks, and 80+ other systems) and use agents to provide faster, cheaper, and better accounting services. \n\n![uploaded image](/media/?type=post&id=110394&key=user_uploads/983067/43e1b7ba-2178-44e3-a001-690ad7b41766)\n\nThere is zero software for customers to use. Just integrate and we’ll close your books for you.\n\nIn the last three months, we went from an idea to partnering with Series B-D companies representing over $1B in enterprise value.\n\n**Team:**\n\n* [Joanathan McIntosh](https://www.linkedin.com/in/joanathanlim/) - Repeat VC-backed founder, built and exited an accounting/ERP platform (45k businesses in 4 years). Ex-Merrill Lynch and StreetLight Data (AI/ML for US Government)\n* [Philip Moniaga](https://www.linkedin.com/in/philipmoniaga/) - Repeat YC founder (W22). Built financial infra at Bloomberg and [Booking.com](http://Booking.com). CTO of a trading platform ($1B+ in volume).\n\n![uploaded image](/media/?type=post&id=110394&key=user_uploads/983067/0bf2a219-f426-444c-a7c2-dda2bbd7e6e7)\n\n**Ask - intros to:**\n\n* CFOs, Controllers, and VPs of Finance at companies with $40M+ in revenue or funding.\n* Fractional CFOs, accounting and advisory firms\n* **$5k referral or 1 US Open ticket for every closed customer.**\n\n[joan@thebillow.ai](mailto:joan@thebillow.ai) | [thebillow.ai](http://thebillow.ai)", "url": "https://www.ycombinator.com/launches/SiY-billow-killing-deloitte-with-ai", "votes": 70, "created_at": "2026-08-13T19:00:00.141Z" } ] }, { "id": 33607, "name": "Atomarine", "slug": "atomarine", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/6485491a38d727a14e926be75b62003c9f8fccff.png", "website": "https://atomarine.co", "all_locations": "San Francisco, CA, USA", "long_description": "We are building nuclear power delivery ships to power our floating compute barges at sea. Our systems are fully standalone, have better cooling, and avoid all the resource competition that leads to local community opposition. Our centralized manufacturing allows us to deploy anywhere on Earth, four times faster than on land.", "one_liner": "Floating nuclear powered data centers at sea.", "team_size": 2, "industry": "Industrials", "subindustry": "Industrials -> Energy", "launched_at": 1784997775, "tags": [ "Hard Tech", "Energy" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "Industrials", "Energy" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/atomarine", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/atomarine.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/atomarine", "x_url": "https://x.com/Atomarine_co", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/6485491a38d727a14e926be75b62003c9f8fccff.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Tyler Bosmeny", "url": "https://www.ycombinator.com/people/tyler-bosmeny" }, "founders": [ { "user_id": 697566, "full_name": "Dimitris Koutentakis", "title": "Founder/CEO", "founder_bio": "Co-founder & CEO at Atomarine. 4x MIT graduate. Previously Portfolio Management Associate at Arrowstreet Capital ($200B AUM), leading quantitative research on trade execution for $2B daily trades and broker selection across $180B annually. At Amazon, designed MILP optimization for electric fleet charging, quantifying $193M CapEx avoidance. MIT undergraduate, MEng in EECS, MS in Mech Eng, MBA from Sloan.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/dkoutentakis/", "x_url": "https://x.com/_dkout", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/ce898826845045bb5604021ca44fae86ccf728df.jpg" }, { "user_id": 3640819, "full_name": "Emile Germonpre", "title": "Founder/CTO", "founder_bio": "Co-founder at Atomarine (YC S26) || PhD in Nuclear Engineering + MSc in Chemical Engineering || Fulbright Scholar + Valedictorian of Ghent University's Engineering School", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/emile-germonpr%C3%A9-766045245/", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/e8bf85b0bc0d2e7dce9804086a8546a8d7738964.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 107426, "title": "Atomarine: Floating Nuclear Data Centers", "tagline": "Floating Nuclear Powered Data Centers At Sea", "body": "\\*\\*We're building floating data centers on barges powered by nuclear ships at sea.\\*\\*\n\n\n\n![uploaded image](/media/?type=post&id=107426&key=user_uploads/697566/5530961e-115c-4010-b749-b4de1a939afa)\n\n\n\\\n**TL;DR:** We move data center campuses offshore, away from communities, onto standardized barges we build on a single manufacturing line. That gets us near-free cooling, no interconnection queue, no local opposition, and 2x faster deployment. First gas-powered pilot ready by 2028.\n\nWhy now: AI has created unprecedented demand for compute and power that cannot be satisfied onshore.\n\n* Grid interconnection timelines are **approaching a decade**.\n* Data centers drive up the local cost of electricity, rent and land, and make water scarce. Community opposition rising from those reasons has **killed more than half of planned projects** since 2025.\n\nWhat we're building: standardized, plug-and-play data center barges, built on one manufacturing line and towed to site. Moving offshore gives us:\n\n* **No interconnection queue and no community impact** because our system operates standalone.\n* More **scalable architecture** that offers **faster deployment** and tighter-controlled deliverability due to centralized construction and repeated site licensing\n* **Lower energy cost** due to **less capacity overbuild** than onshore behind-the-meter and **cheaper financing** for repeatable manufacturing compared to on-site construction (financing = \\~2/3 of the lifetime cost of a nuclear project)\n* Virtually **free cooling** from seawater.\n\nTraction: We already have early partnerships in place across the two hardest parts of this build:\n\n* **Offshore design & construction:** working with established offshore barge designer/builders.\n* **Nuclear power:** partnerships with multiple SMR vendors.\n\nWhy us: We're a two-founder team out of MIT with the exact mix this needs.\n\n* [@Emile Germonpre](https://bookface.ycombinator.com/user/3640819) has a PhD in nuclear engineering.\n* [@Dimitris Koutentakis](https://bookface.ycombinator.com/user/697566) has a BS/MEng in EECS, an MS in Mechanical Engineering, and an MBA, all from MIT, with deep connections to the Greek shipping industry.\n\n**Our ask:** We're looking for intros to decision-makers in compute infrastructure or data center development we can sell to, VPs of Data Center Construction, Directors of Infrastructure, and hyperscaler / neocloud teams. If that's you or someone you know, reach out at [dimitris@atomarine.co](mailto:dimitris@atomarine.co) & [emile@atomarine.co](mailto:emile@atomarine.co).\n\n![uploaded image](/media/?type=post&id=107426&key=user_uploads/697566/f5ce89e6-1e1b-485f-8b35-0e8c5298287e)\n\n", "url": "https://www.ycombinator.com/launches/Rwg-atomarine-floating-nuclear-data-centers", "votes": 24, "created_at": "2026-07-26T20:34:19.952Z" } ] }, { "id": 33639, "name": "Magma", "slug": "magma", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/9b261522e21a27147c25d66eb0b126f8b62e5295.png", "website": "https://magmahq.ai", "all_locations": "San Francisco, CA, USA", "long_description": "We enable companies monetize their agents' traces.", "one_liner": "Monetize your agent's traces.", "team_size": 1, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1786606834, "tags": [ "Reinforcement Learning", "Data Engineering", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/magma", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/magma.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/magma-hq", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/9b261522e21a27147c25d66eb0b126f8b62e5295.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Harshita Arora", "url": "https://www.ycombinator.com/people/harshita-arora" }, "founders": [ { "user_id": 3107607, "full_name": "Ramtin Shahzanian", "title": "Founder/CEO", "founder_bio": "failing my way through.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ramtin-shahzanian/", "x_url": "https://x.com/42Ramtin", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b58098b0fa3c418c52382d092d86f1801136d50d.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 110488, "title": "Magma - AI labs will pay you to learn from your agent's mistakes", "tagline": "We enable any company monetize its agents' traces!", "body": "**TL;DR:** Magma enables companies building AI agents to monetize their data in a few clicks.\n\n\n\n**❌ The problem**\n\nIf a product uses AI agents, every interaction creates valuable data. That could be a user correcting an agent’s mistake, or an agent discovering a new way to complete a workflow. That real-world data currently sits unused, while the best alternative is human annotators inventing hypothetical scenarios.\n\n**✅ What we do**\n\nMagma turns that data into recurring revenue for companies.\n\nCompanies share their agents’ traces in a few clicks. Access is read-only and encrypted, and nothing is shared without approval. Magma anonymizes the data, turns it into complex training environments, and handles licensing to AI labs. Every time the data is licensed, the company gets paid.\n\n**👋 Who I am**\n\nI’ve previously built the first Persian OCR model and legal LLM for my startup in Iran, formulated new transformer-based ML architectures in Oxford, and created a dev tool context management for agents that worked so well I decided to repurpose it to extract training data from traces (now Magma).\n\n![uploaded image](/media/?type=post&id=110488&key=user_uploads/3107607/07bc99c4-63f2-4235-870e-9b97977f8c43)\n\n**🧐 But why?**\n\nWe believe no two AI agents on the planet should be corrected for the same mistake. This is why we're building infrastructure to turn the experience of any AI agent into generalizable learning signals to improve any other AI agent.\n\n**🫡 Our ask**\n\nDo you know anyone at a frontier lab looking for RL environments built from real production workflows, or at an enterprise in need of fine-tuning their models on niche tasks? Let us know and we can help: [ramtin@magmahq.ai](mailto:ramtin@magmahq.ai)\n\nAlso, you can start monetizing your agents’ traces at: [magmahq.ai](http://magmahq.ai)", "url": "https://www.ycombinator.com/launches/Sk4-magma-ai-labs-will-pay-you-to-learn-from-your-agent-s-mistakes", "votes": 37, "created_at": "2026-08-14T01:00:00.832Z" } ] }, { "id": 33673, "name": "Caution", "slug": "caution", "former_names": [ "Caution SEZC" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/a0571039e7dc6cc20c7a181c0b7ffa967bc5c6d3.png", "website": "https://caution.co", "all_locations": "George Town, George Town, Cayman Islands", "long_description": "Caution is a hosting platform for software you can't afford to get hacked. It simplifies and extends confidential compute, and lets developers deploy workloads to secure enclaves in minutes and give customers, auditors, and counterparties cryptographic proof that production is running the exact reviewed source code, build, configuration, and runtime - not an opaque binary controlled by a cloud operator or internal admin.\r\n\r\nConfidential computing hardware exists, but the tooling is too fragmented and incomplete for most teams to use well. Today, attestation often proves only that a binary has a certain hash; it does not prove what source code produced that binary. Building the missing deployment, reproducibility, and verification layer requires specialized security engineering and often locks teams into one cloud or hardware stack. Caution turns that into a simple git-based workflow that builds reproducible enclave images, provisions infrastructure, and lets anyone independently verify what is running.\r\n\r\nWe built Caution after years securing high-risk systems through Distrust, where we worked with hedge funds, custodians, blockchain teams, cloud platforms, and critical grid operators on systems responsible for more than $600B in assets. Again and again, the same problem showed up: teams running mission-critical infrastructure could not reliably prove what was executing in production.\r\n\r\nAs AI, fintech, healthcare, and regulated data collaboration move into the cloud, \"trust us\" is no longer enough. Caution is a general-purpose platform: if it runs on a server, it can run verifiably on Caution. We're starting where trust failures cost the most, but we believe verifiable compute will become the default way software runs, the same way HTTPS became the default way it's served. Infrastructure you can verify, not trust.", "one_liner": "Hosting platform for software you can't afford to have hacked", "team_size": 5, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1785719703, "tags": [ "Cybersecurity", "Cloud Computing", "Infrastructure" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "Cayman Islands", "Latin America", "Remote", "Partly Remote" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/caution", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/caution.json", "year_founded": 2025, "city": "George Town", "country": "KY", "linkedin_url": "https://www.linkedin.com/company/cautionco", "x_url": "https://x.com/CautionHosting", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/a0571039e7dc6cc20c7a181c0b7ffa967bc5c6d3.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 3655130, "full_name": "Anton Livaja", "title": "Founder/CEO", "founder_bio": "Co-founder/CEO of Caution (YC S26). We give teams a hosting platform to run sensitive workloads in secure enclaves and cryptographically prove what code is running in prod. Previously co-founded Distrust, securing systems for hedge funds, custodians, and critical infra responsible for $600B+ in assets. Deep in confidential computing, supply chain security, and applied cryptography. Mission: improve the freedom, security, and privacy of as many people as possible.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/alivaja/", "x_url": "https://x.com/antonlivaja", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/f6dddfa8391b5f9fce21fa7496e2368d59ef6987.jpg" }, { "user_id": 3654857, "full_name": "Ksenia Lesko", "title": "Founder", "founder_bio": "Co-founder at Caution, where we help teams run sensitive workloads in secure enclaves and cryptographically prove what code is running in production. I lead the business and product side; previously ran strategy and ops at Distrust (distrust.co) and led customer education and activation at Ada (ada.cx).", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/ksenialesko/", "x_url": "https://x.com/xenush", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/4ecf3715be1ffcf30eedb062148902b9a1458237.jpg" }, { "user_id": 52788, "full_name": "Lance Vick", "title": "Founder/CTO", "founder_bio": "Technology sovereignty maximalist. Lead security architect at Caution. Former security lead at Turnkey, BitGo, Unit 410, Pebble. Founded Stagex Linux distribution, and #! decentralized hackerspace. Experienced in Linux hardening, privacy, software engineering, security auditing, cryptographic key management, tamper evidence, HSMs, and supply chain security. Frequently internet infamous for exposing security negligence. Prolific mechanical puzzle collector, lockpicker, and hardware hacker.", "is_active": true, "linkedin_url": "https://lance.dev", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/45728fc907c9915858fe6304bbdb25700f05d67b.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 108800, "title": "Caution: Hosting platform for software you can't afford to have hacked", "tagline": "Deploy to secure enclaves in minutes, isolate sensitive workloads, cryptographically prove what code is running.", "body": "Hey everyone! We're [Anton Livaja](https://www.linkedin.com/in/alivaja/), [Lance Vick](https://lance.dev), and [Ksenia Lesko](https://www.linkedin.com/in/ksenialesko/) from [Caution](https://caution.co).\n\n**TL;DR:** Caution is a hosting platform for software you can't afford to have hacked. Your software and data Istay protected even if your CI/CD pipeline, host infrastructure, or a developer laptop gets compromised. You go from source to a verifiable deployment in three commands. \n\n**Built for:** payment processors, PHI processors, AI agents, policy and fraud engines, secrets management, and other sensitive workloads.\n\n[https://drive.google.com/file/d/1u-l2TSBQO2t5dbhBsatCkZf5earCTyDD](https://drive.google.com/file/d/1u-l2TSBQO2t5dbhBsatCkZf5earCTyDD/view?usp=sharing)\n\n### **Why we built this**\n\nOn February 21, 2025, attackers stole **$1.5 billion** worth of Ethereum from Bybit, the largest digital heist in history. They never breached Bybit's own infrastructure: North Korea's Lazarus Group compromised a single developer machine at Safe{Wallet}, Bybit's multisig provider, and injected malicious JavaScript into the signing interface. Bybit's team approved what looked like a routine transfer and handed over a cold wallet to the attackers. The code everyone trusted was not the code that was running.\n\nThis keeps happening because nearly all software runs on blind trust:\n\n* trust in the cloud host\n* trust in production admins\n* trust in CI/CD\n* trust that the code you reviewed is the code actually running\n\nIf any part of that chain is compromised, an attacker can read your secrets, tamper with your business logic, or quietly run code your team never reviewed. It's the same failure behind SolarWinds and Codecov: tampered code arriving through a trusted channel, passing every check. Caution exists to shrink that attack surface and make the most sensitive systems verifiable.\n\n### **Caution proves exactly what's running on your server**\n\nCryptographically, and in real time. Deploying into secure enclaves normally takes a specialist team and months of integration. Caution automates the entire flow: source to verifiable deployment in minutes, with three commands:\n\n![uploaded image](/media/?type=post&id=108800&key=user_uploads/3654857/f3161435-2a97-4207-915a-96375a72be82)\n\n**🎥** [Watch Anton walk through the platform](https://caution.co/platform-tour.html)\n\nUnder the hood, **Caution is a verifiable confidential compute platform:** hardware-backed cryptography isolates your workloads, preserves data privacy, and proves what software is running on a server in real time. The entire platform is open source, so verification never requires trusting us. In practice, this eliminates entire classes of attacks out of the box:\n\n* **Build & CI/CD tampering**: reproducible, multi-party signed builds mean injected code can't pass as a legitimate release.\n* **Infrastructure & insider compromise**: a compromised cloud account, admin, or laptop can't silently swap what's running.\n* **Runtime data theft**: data is processed in hardware-isolated enclaves, unreadable even to the host, the cloud provider, or us.\n\nCaution guards against everything but first- or third-party code bugs: if the release you approve has a flaw, that's what runs. Nothing can be added, no fix can be stripped out, and every patch is provably live.\n\n### **Why now**\n\nWith advances in AI, attacks that used to require rare, expensive expertise are now cheap and scalable, and the gap between a vulnerability existing and being exploited is collapsing toward zero. Reactive patching was already losing. It can't win a race against automated attackers.\n\nIn the post-Mythos world, the security model has to flip. Infrastructure can no longer assume it won't be compromised. It has to be built to stay secure even when the host, the pipeline, or individuals with privileged access are compromised, and to prove, cryptographically, that it has. That's what Caution does.\n\nCaution is a general-purpose platform: if it runs on CPU or GPU, it can run verifiably on Caution. We're starting where trust failures cost the most, but we believe verifiable compute will become the default way software runs, the same way HTTPS became the default way it's served.\n\n### **Why us**\n\n[Anton](https://www.linkedin.com/in/alivaja/) and [Lance](https://lance.dev) have spent their careers securing some of the most security-critical systems in the world: BitGo, Ledn, Turnkey, OpenZeppelin, and Unit 410 (later acquired by Coinbase). Across hundreds of engagements, from hedge funds to critical grid operators, the systems we've helped secure hold more than **$600B** in digital assets. [Ksenia](https://www.linkedin.com/in/ksenialesko/) spent years helping enterprises adopt new technology, and now leads go-to-market, customer education, and operations at Caution, turning deep security engineering into a product companies can understand, buy, and use.\n\nWe've been the people companies call in when this class of attack hits. Teams needed stronger guarantees against supply chain and runtime attacks, but existing solutions were too hard, too bespoke, or too expensive. So we built the platform we kept wishing existed.\n\n### **Traction**\n\nCaution is in production and onboarding customers spanning crypto infrastructure, blockchain labs, digital asset lending, consumer hardware, post-quantum security, and private AI inference.\n\nWe support AWS Nitro Enclaves today. Intel TDX, AMD SEV-SNP, TPM 2.0, NVIDIA Confidential Computing, and broader cloud support is in active development and coming soon.\n\n### **Our ask**\n\nIf you're running software where sensitive data must stay private or business logic can't be tampered with, we'd love to help you level up your security. [Book 20 minutes with us](https://caution.co/contact.html) or email us at [founders@caution.co](mailto:founders@caution.co).\n\nWe'd especially love to hear from you if you're a:\n\n* security team that wants proof, not promises, from your infrastructure\n* fintech running payments, policy, or fraud engines\n* crypto custodian, exchange, or wallet team that can't afford another Bybit\n* healthcare or AI company processing PHI or other regulated data\n* team giving AI agents production credentials and losing sleep over it\n* company using AI inference with sensitive data\n\nYou can also learn more on our [blog](https://caution.co/blog.html), and follow us on [LinkedIn](https://www.linkedin.com/company/cautionco/) and [X](https://x.com/CautionHosting) for updates.", "url": "https://www.ycombinator.com/launches/SIq-caution-hosting-platform-for-software-you-can-t-afford-to-have-hacked", "votes": 59, "created_at": "2026-08-04T17:02:42.242Z" } ] }, { "id": 33738, "name": "CoArena", "slug": "coarena", "former_names": [ "Coasty" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/1543623e6506eaf893b1c43e614688a6b2a32bda.png", "website": "http://coarena.ai", "all_locations": "San Francisco, CA, USA", "long_description": "CoArena is a live arena where anyone can use the world's top computer-use models racing two of them on the same real computer task and judging which one did it better. Every battle becomes something the AI labs can't build themselves, an honest test of their agents on real work, and the data to make them better.", "one_liner": "The biggest crowdsourced benchmark for Computer-Use", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Engineering, Product and Design", "launched_at": 1782294083, "tags": [ "AIOps", "Machine Learning", "Reinforcement Learning", "Data Labeling", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Engineering, Product and Design" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/coarena", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/coarena.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/coastyai", "x_url": "https://x.com/coastyai", "fb_url": null, "cb_url": "https://www.crunchbase.com/organization/coasty", "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/1543623e6506eaf893b1c43e614688a6b2a32bda.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/cfec1f74a71f276cfe85bcc102647baf48c51fff.png" ], "primary_partner": { "name": "Brad Flora", "url": "https://www.ycombinator.com/people/brad-flora" }, "founders": [ { "user_id": 1194042, "full_name": "Prateek Jannu", "title": "Founder/CEO", "founder_bio": "Prateek Jannu, Co-founder & CEO - Stanford (MS. EE) and Purdue (BE. CE). Invented Solo-Synth-GAN, which generates labeled training data from a single example. Built Coasty to #1 on the OSWorld benchmark (82%), beating Agent S3, UiPath, Kimi K2.5, and Claude Sonnet 4.5. Now building CoArena.", "is_active": true, "linkedin_url": "https://linkedin.com/in/prateekjannu", "x_url": "https://x.com/prateekjannu", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/26ed45c60c34293a10373502e615a8c6499d356f.jpg" }, { "user_id": 2457255, "full_name": "Nitish Kovuru", "title": "Founder", "founder_bio": "Co-founder and CTO/COO of Coarena, launched Coasty prev, SOTA CUA framework on OSWorld at 83% accuracy. Did CS at Columbia (Vision track) and hold a BS in Computer Engineering from Purdue, where I TA'd a graduate-level AI course as an undergrad and later went on to build enterprise sales automation agentic systems for my previous company.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/nitish-kovuru-2423351b4", "x_url": "https://x.com/llmhubdev", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/29f588d1c332c655eee8ea69c5e1c27e9ec4fce9.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111689, "title": "Coarena - The biggest crowdsourced benchmark for Computer-Use", "tagline": "We build the most advanced and biggest benchmark for computer use", "body": "We are officially launching CoArena (YC S26). Released 2 weeks ago, users have doubled every week since, and the arena went from $0 to $60,000 in revenue since. It's a live arena where the world's top Computer-Use models race to finish the same real computer task, and real people judge which one did it better, blind.\n\n\n\n![uploaded image](/media/?type=post&id=111689&key=user_uploads/1194042/43ddaa95-1f50-4b30-b0a4-9b085433ec3c)\n\n**The problem**\n\nEvery AI lab says their agent is the best at using a computer, and every claim points at a benchmark score. We know those scores well: earlier this year we built the #1 agent on OSWorld (82%). That's exactly how we learned the scores mean very little. Benchmarks are static, agents memorize them, and they say nothing about the messy tasks people actually need done.\n\n**What we built**\n\nOn CoArena, anyone can type a real task (\"find me the cheapest flight,\" \"assemble this expense report\"), watch two anonymous frontier agents race it live, and vote on the winner. Names are revealed only after you vote, so the judging is blind. Every battle becomes a fresh eval no model has ever seen, and the failure cases labs never catch: prompt injection, spending money when it should ask, silent side-effects.\n\n**What the arena has already shown**\n\n\\- Only \\~66% of agent runs finish their task at all 1 in 3 dies to a failure before any score exists\n\n\\- On identical tasks, the best frontier agent completes \\~87% of runs; the worst completes \\~42%\n\n\\- One major frontier model times out in roughly a third of its runs\n\n\\- In \\~7% of battles, human judges ruled that BOTH agents failed\n\n\\- The #1 spot on our leaderboard changed hands three times in three days, real tasks don't saturate\n\n**What you get**\n\nIt's completely free, including models you can't use anywhere else (Fable, Sol, and unreleased previews). Run your real errands with frontier agents, keep whatever they produce, and find out which model is actually best at YOUR kind of work, not the benchmark's.\n\n**Our asks**\n\n1\\. Use it, it's free. Bring your daily work: the flights, the spreadsheets, the research. Two frontier agents will race it, you keep the result, and your vote makes the leaderboard more honest: \n\n2\\. If you're an AI lab working on computer use or vision, or a robotics company building on vision, we'd love to talk. We can send you sample evals and data today: [founders@coasty.ai](mailto:founders@coasty.ai)", "url": "https://www.ycombinator.com/launches/T3R-coarena-the-biggest-crowdsourced-benchmark-for-computer-use", "votes": 59, "created_at": "2026-08-22T19:00:00.303Z" }, { "id": 105157, "title": "Coasty: The computer-use agent that doesn't break on real software", "tagline": "Coasty helps developers kill RPA.", "body": "**TL;DR: Coasty gives developers a computer-use agent that reliably automates back-office work in legacy desktop software where RPA scripts and other CUAs fail. It’s achieved 82.81% on OSWorld and it’s great at automating workflows where quiet mistakes cost the most: regulatory filings, patient records, payments, and gives developers precise control and observability over its work.**\n\n\\\n(_The portals and the data in it, are mockups_)\n\n**Replayable audit trail (real run):**  \n\n**OSWorld submission (Coasty scored 82.81% on OSWorld Verified, confirmed by the OSWorld team, with our latest internal run at 85.60%.):**[ https://github.com/coasty-ai/coasty-osworld](https://github.com/coasty-ai/coasty-osworld)\\\n\\\n**The Problem**\n\nMost back-office work doesn't happen in clean web apps. It happens in legacy desktop software, full of surprise pop-ups, wandering menus, and impenetrable UIs.\n\nToday, developers automating this back-office “real” work are stuck choosing between brittle RPA scripts and CUAs that demo beautifully and score high but break on contact.\n\nThe problem isn't getting an agent to click. It's trusting it when a silent mistake gets someone fired.\n\n### **Why Coasty is reliable where others aren't**\n\n**It recovers mid-task.** Coasty catches its errors and fixes them mid-run. If it’s crunching on a 20-step workflow and makes a misclick at step 3, it will recover and start from the error. First-pass accuracy is vanity. Recovery keeps a run alive.\n\n**It survives messy software.** Pop-ups, modals that steal focus, legacy desktop apps. Coasty doesn’t record clicks, it reads the screen. So when the UI changes overnight and every RPA script breaks, Coasty looks at the new screen and keeps going.\n\n**It knows when it's done.** An agent that thinks it finished but didn't is worse than one that fails loudly. Coasty verifies its own output and leaves a timestamped audit trail you can replay.\n\n**It swarms.** One task across forty state portals, each with its own ancient interface, all at once. Forty agents, one verified result set. What takes a team a week, is done overnight.\n\n### **Traction**\n\nCoasty has been verified by the OSWorld team and will be published soon, but also another internal benchmark we published (85.60%): < >.\n\nWe have paying customers on four continents, a few thousand active users, and we have actually grown 24x since Claude Cowork came out.\n\n### **Why us**\n\nWe started working on this idea when models weren’t good enough for computer use (last summer), so we’ve built frameworks to test and run the perfect combination of models and harnesses to see what works in practice. Prateek worked as an ML engineer at multiple startups and dropped his Stanford admit while I did CS at Columbia and built my previous company’s entire agentic sales automation. We’ve grown with this field and have been working on it even before it got much attention, so we understand where CUAs mess up and how to deal with these messes realistically.\n\n### **The ask**\n\nSend us one workflow your team runs by hand. We’ll tell you straight whether Coasty can run it reliably today. If it looks automatable, we’ll build a working proof of concept on your real software before a long sales process.\n\nWe’re especially looking for intros to operators and executives in accounting, tax, insurance, healthcare, compliance, and data privacy - anywhere teams are still babysitting RPAs, clicking through legacy portals, reconciling records, filing forms, or moving data across old desktop software.\n\nHere's our Calendly:[ https://cal.com/coasty/15min](https://cal.com/coasty/15min)", "url": "https://www.ycombinator.com/launches/RM5-coasty-the-computer-use-agent-that-doesn-t-break-on-real-software", "votes": 12, "created_at": "2026-07-10T17:30:25.886Z" } ] }, { "id": 33805, "name": "Edviro", "slug": "edviro", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c179a28044a3d4a8c3ceaf2a5b8cecbf90f66aa7.png", "website": "https://edviroenergy.com", "all_locations": "San Francisco, CA, USA", "long_description": "Edviro builds AI that operates energy infrastructure. We connect to meters, HVAC systems, batteries, solar, and other energy equipment and continuously handle the work required to keep them efficient: detecting and diagnosing problems to forecasting demand, simulating fixes, creating work orders, coordinating maintenance, and controlling equipment where authorized. Our energy world models learn how each site behaves, choose the highest-value actions, and verify the results against real operational, meter, and billing data.", "one_liner": "AI That Operates Energy Infrastructure", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1782807553, "tags": [ "Deep Learning", "B2B", "Energy" ], "tags_highlighted": [], "top_company": false, "isHiring": true, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/edviro", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/edviro.json", "year_founded": 2025, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/edviro", "x_url": "https://x.com/Edviro_energy", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c179a28044a3d4a8c3ceaf2a5b8cecbf90f66aa7.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 3164154, "full_name": "Hursh Shah", "title": "Founder/CEO", "founder_bio": "19, Founder & CEO at Edviro. Prev. NeurIPS 2024, ChatGPT Lab. Patent pending medtech research. Prev. founded Skyglass, smart glasses startup that got accepted to PearX. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/hursh-shah/", "x_url": "https://x.com/hursheybar2", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/9756e91b633ed2b9de6793c88f0e1daa8ff1d704.jpg" }, { "user_id": 1865110, "full_name": "Tanuj Siripurapu", "title": "Founder", "founder_bio": "19, Founder & CTO @ Edviro (YC S26). Prev. at RTX. Scaled a digital presence agency to 20k revenue in high school. Programming since age 9", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/tanuj-siripurapu-747b58217/", "x_url": "https://x.com/kstanuj", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/96aabcf62bd08eff9921363a085eb3a930988175.jpg" } ], "jobs": [ { "id": 113407, "title": "Member of Technical Staff", "url": "https://www.ycombinator.com/companies/edviro/jobs/PlLCBWC-member-of-technical-staff", "location": "San Francisco, CA, US / Remote (US)", "type": "Full-time", "role": "Engineering", "role_type": "Machine learning", "salary_range": "$160K - $260K", "equity_range": "0.50% - 2.00%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] }, { "id": 113409, "title": "Founding GTM", "url": "https://www.ycombinator.com/companies/edviro/jobs/xuWBSdL-founding-gtm", "location": "San Francisco, CA, US / Remote (US)", "type": "Full-time", "role": "Sales", "role_type": null, "salary_range": "$140K - $170K", "equity_range": "0.50% - 1.50%", "experience": "Any (new grads ok)", "visa": "US citizen/visa only", "skills": [] } ], "news": [], "launches": [ { "id": 112256, "title": "Edviro - AI that operates energy infrastructure", "tagline": "AI to operate energy infrastructure", "body": "# **Edviro: AI that operates energy infrastructure**\n\nHi all,\n\nWe’re Hursh and Tanuj, founders of Edviro.\n\n**TL;DR:** Edviro connects to a facility’s BMS, utility, sensor, equipment, and maintenance systems to build a world model of how the facility behaves. We use that model to detect problems, simulate interventions before anything is changed, and coordinate the best next action. We’re live in **34 facilities**, have identified more than **$400K in avoidable energy costs**, and are now opening our first partnerships with data center operators.\\\n\\\n\n\n## **The problem**\n\nData centers are expected to drive roughly half of U.S. electricity-demand growth through 2030.\n\nThis is usually framed as a problem of producing more electricity. That is only part of it.\n\nPower must be delivered to the right sites, on the timelines data centers are being built, with the reliability their workloads require. Once a facility has that power, its operators still need to extract as much useful compute as possible from every available megawatt without compromising uptime.\n\nInside a data center, power, cooling, controls, equipment, workloads, and maintenance are tightly coupled. A workload shift changes the thermal profile. A cooling-control change affects both energy use and equipment margins. A maintenance decision can alter redundancy and available capacity.\n\nYet the data and workflows used to manage these systems remain fragmented across BMS, DCIM, EPMS, meters, sensors, maintenance software, vendor portals, and spreadsheets.\n\nMost existing systems can display telemetry or raise an alarm. They cannot reliably model how the entire facility will respond before an operator changes a setpoint, shifts a workload, services equipment, or adds capacity.\n\nOperators are left making high-stakes decisions with incomplete system-level context.\n\n## **What Edviro does**\n\nEdviro turns a facility’s operational data into a continuously updated world model.\n\nBy “world model,” we mean a model that estimates the current state of the physical system, learns how its components interact, and predicts how it will respond to a proposed action. \n\nEdviro:\n\n* Integrates data from building controls, utilities, sensors, equipment, and maintenance systems\n* Detects anomalies, energy waste, equipment degradation, and capacity constraints\n* Simulates operational and maintenance interventions before they are deployed\n* Recommends the best action while accounting for cost, reliability, and physical constraints\n* Uses agents to coordinate approvals, work orders, investigations, and follow-through\n* Verifies whether the expected operational and energy impact actually occurred\n\nFor a data center, this could mean determining whether existing power and cooling infrastructure can support additional IT load, testing a control-sequence change before deploying it, identifying which equipment issue is constraining capacity, or prioritizing maintenance based on the capacity and reliability it would recover.\n\nWe initially keep operators in control of every consequential action. As the models are validated and earn trust, more of the operating loop can be automated.\n\n## **Why we started with schools**\n\nEdviro began when I was 17 and working with my high school’s facilities team. I was given a year of district utility bills to analyze, built software to parse them, and discovered roughly **$360,000 in gas overbilling**.\n\nWe found that this was not an isolated problem. Schools operate large portfolios of aging, mixed-vendor infrastructure with fragmented controls, incomplete data, lean facilities teams, and very little room for wasted budget.\n\nThey forced us to build for the messy reality of physical infrastructure rather than clean laboratory data.\n\nToday, Edviro is live in **34 facilities** and has identified more than **$400,000 in avoidable energy costs** for building operators.\n\nA school is obviously not a data center. Data centers have denser instrumentation, much tighter uptime requirements, and more complex coupling between power, cooling, and compute. \n\nWhat transfers is the underlying architecture: connecting heterogeneous operational systems, estimating the state of a physical facility, simulating interventions, coordinating action across teams, and verifying the result. We are now building the data-center-specific models, integrations, and safety constraints directly with operators. Our first relevant research benchmark can be found here: \n\n## **Where we’re going**\n\nThe current generation of facilities software tells operators what has already happened.\n\nWe want Edviro to understand what is happening now, predict what will happen next, and help execute the safest and highest-value response.\n\nAs electricity becomes one of the primary constraints on AI growth, building more infrastructure will not be enough. We also need to operate the infrastructure we already have far more intelligently.\n\nThe long-term goal is not another dashboard. It is infrastructure that can understand itself and increasingly operate itself.\n\n## **Asks**\n\n1. **We’re looking for data center operators to become early design partners.** We are especially interested in facilities dealing with power or cooling constraints, capacity-expansion planning, recurring manual investigations, fragmented controls, or reactive maintenance. We can begin read-only, with operators retaining approval over every action.\n2. **We’re hiring energy-modeling engineers** across thermal systems, power systems, controls, scientific machine learning, and physical simulation.\n\nIf you operate a data center—or know the team responsible for operating one—we’d love to talk.\n\n[**www.edviro.com**](http://www.edviro.com)", "url": "https://www.ycombinator.com/launches/TCa-edviro-ai-that-operates-energy-infrastructure", "votes": 5, "created_at": "2026-08-26T17:00:00.166Z" }, { "id": 104853, "title": "Edviro - Buildings are wasting money daily", "tagline": "Your building is wasting energy daily", "body": "**TL;DR:** Edviro helps facilities managers at large buildings save on their energy bills.\n\nFacilities managers are overwhelmed with dashboards, alarms, and analytics for hundreds of boilers, HVAC units, and heat pumps across their buildings. Edviro continuously monitors and optimizes buildings using energy world models to detect anomalies, simulate interventions, and suggest the best options to the facilities manager. In particular, in busy spaces like offices, we utilise Wi-Fi access points or our own occupancy sensors to reduce HVAC usage significantly. Our tools also allow institutions to maintain detailed service histories and equipment logs and to perform maintenance with AR. To date, we've saved clients six figures in multiple buildings. We’re also piloting with a construction company to conduct baselining and M&V for sites they work with.\n\n**The team:**\n\nWe’re young technical founders. I’ve published at NeurIPS ‘24, conducted patent pending VR Alzheimer’s work, and founded a previous Aviation smart glasses startup that was accepted into PearX. Tanuj has been coding since he was 9, scaled a previous digital agency to $20k in revenue and previously did dev work at RTX.\n\nThis all started when I was working with my own high school’s facilities team. I saw how much waste there was and how fragmented his energy data was despite loads of smart controls and BMS systems. Initially I built a dashboard for him, and the product quickly caught on to more schools in NorCal.\n\n**Our ask:**\n\nWe’re looking for intros to facilities teams at school districts, real estate or data centres.\n\nAlongside that, we’re looking to work with construction teams who need help baselining and doing M&V on their work post-project.\n\nIf you are connected to anyone in this space, would love to chat!\n\n![uploaded image](/media/?type=post&id=104853&key=user_uploads/3164154/9fe82a05-dd5a-4916-94f4-9cf0e8f54a05)\n\n", "url": "https://www.ycombinator.com/launches/RHB-edviro-buildings-are-wasting-money-daily", "votes": 5, "created_at": "2026-07-07T15:02:05.737Z" } ] }, { "id": 33806, "name": "Experiential Labs", "slug": "experiential-labs", "former_names": [ "Resolute Labs" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/716e828c9efd6efe6da64574ff67dd5bfa0c2a25.png", "website": "https://experientiallabs.ai", "all_locations": "San Francisco, CA, USA", "long_description": "We are building an open source gateway for AI budget holders to control model access and turn their traffic into better models that they own. We're a team that has earned an ML PhD at Georgia Tech, scaled AutoGPT to 160k GitHub stars, and led a self-driving simulation team at Waabi.", "one_liner": "Open source AI gateway that turns your traffic into better models", "team_size": 2, "industry": "B2B", "subindustry": "B2B -> Infrastructure", "launched_at": 1782607347, "tags": [ "Reinforcement Learning", "Infrastructure", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B", "Infrastructure" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/experiential-labs", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/experiential-labs.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/experiential-labs/", "x_url": "https://x.com/explabsai", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/716e828c9efd6efe6da64574ff67dd5bfa0c2a25.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Diana Hu", "url": "https://www.ycombinator.com/people/diana-hu" }, "founders": [ { "user_id": 3149739, "full_name": "Kion Fallah", "title": "Founder/CEO", "founder_bio": "Co-founder and CEO of Experiential Labs. Previously Staff Research Scientist at Waabi leading the mixed reality simulation team. PhD in Machine Learning from Georgia Tech focused on synthetic data.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/kfallah", "x_url": "https://x.com/okfallah", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/eb099f2fb8bbfee19779eae1dae81307ea85d07d.jpg" }, { "user_id": 411375, "full_name": "Silen Naihin", "title": "Founder/CTO", "founder_bio": "Built and grew AutoGPT to 160k stars. Did AI for science at the DOE. Previously cofounded Stackwise (was in W24) and various founding AI eng roles. Did continual learning, interp, and evals research.", "is_active": true, "linkedin_url": "https://linkedin.com/in/silen-naihin", "x_url": "https://twitter.com/silennai", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/b68b10165c34092f3bb0c78caccfa5794a625212.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 109456, "title": "Experiential Labs - Turn your AI spend into a cheaper + better model you own", "tagline": "Train a model your company owns, guaranteed to match or beat frontier quality at 50% lower cost.", "body": "**TL;DR: Companies are spending more on AI every month, but almost none of that spending becomes an asset they own. We train a small language model for your company's use cases. You fully own it, and we guarantee the same or better quality as the frontier models you use today at 50% lower cost.**\n\n\n\nCompanies are spending more on AI every month, but almost none of that spending becomes an asset they own.\n\nThe prompts, traces, and use cases are yours, _but the model stays rented._\n\nSo you keep paying frontier prices for repetitive work a smaller, specialized model could do faster, cheaper, and often better\n\nWe built Experiential Labs to change that.\n\n**Here's how it works**\n\nKeep using GPT, Claude, or whatever models are already in your stack.\n\nExperiential Labs uses your traffic to build a simulation of the tasks, failures, and edge cases your agent encounters. You can also upload existing traces to bootstrap the process.\n\nThe platform trains you a model against that simulation, so it gets better at your specific use cases over time.\n\nAs your model improves, it earns more traffic. Hard requests can still go to the frontier. Everything else moves to your faster, cheaper model.\n\nYou own the weights, can run them wherever you want, and keep the profits they create. When a stronger base model comes out, we transfer what your model has learned instead of making you start over.\n\nOn some tasks we've seen up to 97% lower cost and +50% higher quality. Across a full deployment, we guarantee at least 50% lower cost at the same or better quality, backed by an SLA.\n\n**The team**\n\nWe are both AI researchers. Kion built simulation systems for autonomous vehicles at Waabi after earning his ML PhD at Georgia Tech. Silen built and grew AutoGPT to 160,000 GitHub stars and has worked across continual learning, evaluations, interpretability, and AI for science.\n\n## Our ask\n\nIf your company sends significant traffic to GPT, Claude, or other frontier models, we would love to help.\n\nIntros to founders and AI leaders with growing inference bills are also greatly appreciated.\n\n**Email:** [founders@xplabs.ai](mailto:founders@xplabs.ai) \n\n**Schedule a call:** \n\n**Website:** [xplabs.ai](http://xplabs.ai) ", "url": "https://www.ycombinator.com/launches/STQ-experiential-labs-turn-your-ai-spend-into-a-cheaper-better-model-you-own", "votes": 18, "created_at": "2026-08-06T22:00:00.114Z" } ] }, { "id": 33936, "name": "Conifer", "slug": "conifer", "former_names": [], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/1cd2adb8270db80170ab62b4140817892372538a.png", "website": "https://www.conifer.build", "all_locations": "San Francisco, CA, USA", "long_description": "There are thousands of models, providers, and harnesses, each with its own pricing and strengths, leaving an overwhelming number of choices when it comes to managing token spend and intelligence. Conifer uses intelligent routing and orchestration logic to handle the full path of each query: model selection, provider choice, and cache management. It all runs inside your existing harness, so tools like Claude Code and Codex work without changing your setup. By centralizing where inference occurs, Conifer lets companies save on their inference bill while leaving teams free to focus on what they are building.", "one_liner": "Least cost routing system to reduce 70%+ token spend", "team_size": 4, "industry": "B2B", "subindustry": "B2B", "launched_at": 1783524440, "tags": [ "Developer Tools", "B2B", "Security", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/conifer", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/conifer.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/coniferbuild/posts/?feedView=all", "x_url": "https://x.com/coniferbuild", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/1cd2adb8270db80170ab62b4140817892372538a.png", "app_video_url": null, "dday_video_url": null, "company_photos": [], "primary_partner": { "name": "Gustaf Alstromer", "url": "https://www.ycombinator.com/people/gustaf-alstromer" }, "founders": [ { "user_id": 2679506, "full_name": "Charles Muehlberger", "title": "Founder/CTO", "founder_bio": "Founder at Conifer. Engineering researcher at Princeton, accelerating multi-modal inference on edge devices and designing physics simulation software for coupled laser reservoirs and machine learning with physical neural networks. Previously built custom edge AI devices for RF-based brain injury modeling at DoW, and helped engineer the world record holding fastest electric speedboat at PES.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/charlesmuehl/", "x_url": "https://x.com/charles_v11", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/409f2cd05975613a17cb1fd3e79e3a7f552feca0.jpg" }, { "user_id": 3664042, "full_name": "Michael Jeffords", "title": "Founder/CGO", "founder_bio": "Co-founder at Conifer (YC S26). Previously co-founded The HeartCheck Foundation, scaling free hypertension screenings to 20+ barbershops and over 10,000 individuals. Engineered an ML-driven computer vision pipeline using pose estimation for early-onset ALS and Parkinson's detection. Former clinical researcher at the Dell Med Institute of Vascular Surgery and UTHSC, specializing in hematopoetic stem cell transplantation, achieving infiltration levels >90% in mouse models. ", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/michael-bryan-jeffords/", "x_url": "https://x.com/michaelbjeff", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/58eb17d9512cddea2d8c88e01b06698b9d88a191.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 105897, "title": "Conifer: Local-first least cost routing system", "tagline": "Stop paying cloud prices for every token. We route 80% of requests to your hardware at no cost", "body": "Hi all 👋, We’re Michael and Charles, founders of Conifer.\n\n**TLDR:** Conifer is one interface for both local and cloud models, available today as **Juniper**, our consumer-facing app. It routes each request to the most cost-effective option, beginning with your own hardware. Most requests are processed locally without API fees, reducing paid token volume by up to 80%.\n\n**📺 Launch Video:**[ https://www.youtube.com/watch?v=QLVNISmtet0](https://www.youtube.com/watch?v=QLVNISmtet0)\n\n**❓ The problem**: Model labs direct every request to the cloud, with no incentive to route your queries to a cheaper competitor or a free, on-device model. That may benefit them, but it’s unnecessarily costly for you. Whether it’s a simple typo fix or a complex architecture problem, the request is sent to a data center, exposing your data to third parties and increasing your monthly spend. This quickly adds up for teams running coding agents, customer support, and other high-volume workloads. When handling financials, patient records, or customer data, teams must also trust that a cloud provider’s security is as robust as promised.\n\n![uploaded image](/media/?type=post&id=105897&key=user_uploads/3664042/6dbc9654-b4df-453d-a02e-3ea71302316d)\n\n**🔨 What Conifer does:** Conifer routes every request from the bottom up:\n\nTier 0: Your own hardware, with $0 in API fees\n\nTier 1: An efficient cloud model\n\nTier 2: A frontier model for the most demanding queries\n\nThe majority of requests are processed locally, ensuring that cloud pricing applies only when actually necessary, drastically reducing paid token volume. Additionally, Conifer consolidates disparate subscriptions, dashboards, and API keys within a single unified interface.\n\nFor sensitive workloads 🕵️, **local-only mode** disables cloud routing entirely, so all conversations and data remain completely on-device.\n\n🏃‍♂️ **Ensuring the local tier is fast enough:** This routing only works if the local tier is reliable, so we built Conifer's inference engine from the ground up in Rust. On Apple Silicon, decode speeds reach up to 60% faster than llama.cpp.\n\n![uploaded image](/media/?type=post&id=105897&key=user_uploads/3664042/79b464dc-667b-48af-a1ba-5cdd089b1dc8)\n\nOn-device operation with Conifer is now the default starting point for inference, rather than a feature limited to hobbyist use.\n\n**Asks:**\n\n1. Try Juniper with a one-click download @[ conifer.build](http://conifer.build), or reach out to us at [contact@conifer.build](mailto:contact@conifer.build) to have a model running in under two minutes.\n2. Introduce us to teams with high monthly token spend. If teams are incurring significant API costs and want to optimize unit economics without rebuilding their stack, we would like to connect.", "url": "https://www.ycombinator.com/launches/RY1-conifer-local-first-least-cost-routing-system", "votes": 10, "created_at": "2026-07-15T22:29:01.603Z" } ] }, { "id": 33969, "name": "OpenTag", "slug": "opentag", "former_names": [ "Gini Agent", "Gini", "OpenTag", "Open Tag" ], "small_logo_thumb_url": "https://bookface-images.s3.amazonaws.com/small_logos/c893158d06fb0a2865c30711a691fd486b8a7dcf.png", "website": "https://www.tryopentag.com/", "all_locations": "San Francisco, CA, USA", "long_description": "OpenTag is an AI coworker in your Slack with full context on your company. It answers questions and takes real work off your plate. It also keeps your wiki up to date automatically, so knowledge never goes stale.", "one_liner": "Model Agnostic AI coworker in Slack", "team_size": 3, "industry": "B2B", "subindustry": "B2B", "launched_at": 1787093671, "tags": [ "AIOps", "B2B", "Workflow Automation", "Enterprise Software", "AI" ], "tags_highlighted": [], "top_company": false, "isHiring": false, "nonprofit": false, "batch": "Summer 2026", "status": "Active", "industries": [ "B2B" ], "regions": [ "United States of America", "America / Canada" ], "stage": "Early", "app_video_public": false, "demo_day_video_public": false, "app_answers": null, "question_answers": null, "url": "https://www.ycombinator.com/companies/opentag", "api": "https://devasheeshg.github.io/yc-api/batches/summer-2026/opentag.json", "year_founded": 2026, "city": "San Francisco", "country": "US", "linkedin_url": "https://www.linkedin.com/company/109881466", "x_url": "", "fb_url": null, "cb_url": null, "github_url": null, "logo_url": "https://bookface-images.s3.amazonaws.com/small_logos/c893158d06fb0a2865c30711a691fd486b8a7dcf.png", "app_video_url": null, "dday_video_url": null, "company_photos": [ "https://bookface-images.s3.us-west-2.amazonaws.com/attachments/790aa3ec5de0bba9ece0a6519e8daba6f10a2cff.png" ], "primary_partner": { "name": "Jared Friedman", "url": "https://www.ycombinator.com/people/jared-friedman" }, "founders": [ { "user_id": 802295, "full_name": "Tony Kam", "title": "Founder/CEO", "founder_bio": "Co-founder and CEO at OpenTag. Previously at Tesla and founded Lilac Labs (S24). Studied EECS at Berkeley.", "is_active": true, "linkedin_url": "https://linkedin.com/in/tonyshekkam", "x_url": "https://x.com/tonykamshek", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/79ba45d01b5c5889e6430069721b3a279fa04e29.jpg" }, { "user_id": 3880379, "full_name": "Wilson Nguyen", "title": "Founder", "founder_bio": "Co-founder at OpenTag. Previously founding engineer at Lilac Labs (YC S24). Studied Computer Science at UC Berkeley.", "is_active": true, "linkedin_url": "https://www.linkedin.com/in/nguyenwilson/", "x_url": "https://x.com/potatoparser", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/19a40eb53a4be1df58ecf47644f1c5154bd9ff7e.jpg" }, { "user_id": 1378729, "full_name": "Shelden Shi", "title": "Founder/CTO", "founder_bio": "Co-Founder and CTO at OpenTag. Previously founded Lilac Labs (S24). Studied CS and CogSci at UC Berkeley.", "is_active": true, "linkedin_url": "https://linkedin.com/in/sheldenshi", "x_url": "", "avatar_thumb_url": "https://bookface-images.s3.us-west-2.amazonaws.com/avatars/92d85249cf52d37811253ebfde1164c62804c115.jpg" } ], "jobs": [], "news": [], "launches": [ { "id": 111999, "title": "OpenTag - Model Agnostic Claude Tag", "tagline": "AI Super Coworker in Slack.", "body": "Hey everyone. You’ve might of heard of Open AI, Open Router, OpenCode, OpenWork….\n\nIntroducing [OpenTag](https://www.tryopentag.com/?utm_source=yc&utm_medium=referral&utm_campaign=yc-launch&utm_content=launch-page) - the model agnostic Claude Tag.\n\nWe started early access a 2 weeks ago and are live with 10 teams.\n\nIf you've seen Claude Tag, you know the interface: tag an AI in Slack, hand it a task, get the result in the thread. We think that's the right interface. We built OpenTag because we think the multiplayer harness should be model agnostic. When a better model ships, your coworker gets better the same day, benefiting from the performance and cost improvements.\n\nOpenTag is an AI coworker that lives inside of Slack. You invite it into the channels you want. It picks up how the company runs, how decisions are made, who owns what. It helps the team coordinate and builds automations for workflows that it sees as repetitive. It asks for tools and visibility the way an A-player hire does, and it tries to grow its own scope and agency. It starts as an intern and tries to earn the COO job. As it works, it writes a company wiki of how you actually run. That wiki compounds as a company brain, so OpenTag only gets sharper with time. Throw it anything. It does its best, and it asks for help when it’s blocked. The result lands in the thread, with sources, where the whole team can see it.\n\n**Four parts, one coworker:**\n\n1. **Ambient and proactive agents -** properly sandboxed with its own computer and browser, so you can trust them with actual tasks instead of babysitting chat windows.\n2. **A wiki that maintains itself —** no more stale handbooks. OpenTag writes down how the company actually runs and keeps it current. Teammates can read it. Agents can read it.\n3. **Automations that suggest themselves -** OpenTag watches how your team works and surfaces the repetitive stuff worth automating, so you're not designing workflows from a blank page.\n4. **Built for multiplayer** - OpenTag is designed with proper permissioning that matches how your team operates.\n\n**2 minutes to start.** Add OpenTag to Slack, invite it to the channels where work happens, give it a job. It’ll ask for the tools it needs (whatever the job touches). It only reads the channels you invited it to.\n\n**Built by software engineers who’ve shipped products used by millions.**\n\nThe three of us worked together at Lilac Labs (YC S24) and all went to Berkeley. Tony (CEO) was software at Tesla and Intel. Shelden (CTO) was at Flatiron Health, Roche, and Certik (cyber security). Wilson (Chief Engineer) was at Hewlett Packard Enterprise and Vestible, a regulated securities platform.\n\nOpenTag is the coworker we wanted at our last company, when we were the ones running every job.\n\n**Try OpenTag today**\n\nTry it at [www.tryopentag.com](http://www.tryopentag.com)\n\nTake 30 minutes with me: [book a demo.](https://cal.com/tony-kam-ympx2p/intro?overlayCalendar=true)\n\nWould love to hear your thoughts and feedback. You can email me at [tony@tryopentag.com](mailto:tony@tryopentag.com)\n\n![uploaded image](/media/?type=post&id=111999&key=user_uploads/802295/8d1c1925-f213-405e-bfd4-f9b9effd0095)\n\n", "url": "https://www.ycombinator.com/launches/T8R-opentag-model-agnostic-claude-tag", "votes": 6, "created_at": "2026-08-25T22:00:00.175Z" }, { "id": 111423, "title": "Gini: AI coworker that lives in Slack", "tagline": "Does the work in your channels and turns what it learns into a company brain.", "body": "Hi everyone, [Shelden](https://www.linkedin.com/in/sheldenshi), [Wilson](https://www.linkedin.com/in/nguyenwilson), and I ([Tony](https://www.linkedin.com/in/tonyshekkam/)) are cofounders of [Gini](https://ginicomputer.com/?utm_source=yc&utm_medium=referral&utm_campaign=yc-launch&utm_content=launch-page).\n\n[Gini](https://ginicomputer.com/?utm_source=yc&utm_medium=referral&utm_campaign=yc-launch&utm_content=launch-page) is a model agnostic proactive AI coworker that lives inside of Slack. You invite it into the channels you want. It picks up how the company runs, how decisions are made, who owns what. It helps the team coordinate and builds automations for workflows that it sees as repetitive. It asks for tools and visibility the way an A-player hire does, and it tries to grow its own scope and agency. It starts as an intern and tries to earn the COO job. As it works, it writes a company wiki of how you actually run. That wiki compounds as a company brain, so Gini only gets sharper with time. Throw it anything. It does its best, and it asks for help when it’s blocked. The result lands in the thread, with sources, where the whole team can see it.\n\n\n\n**Four parts, one coworker:**\n\n1. **Ambient and proactive agents -** properly sandboxed with its own computer and browser, so you can trust them with actual tasks instead of babysitting chat windows.\n2. **A wiki that maintains itself —** no more stale handbooks. Gini writes down how the company actually runs and keeps it current. Teammates can read it. Agents can read it.\n3. **Automations that suggest themselves -** Gini watches how your team works and surfaces the repetitive stuff worth automating, so you're not designing workflows from a blank page.\n4. **Built for multiplayer** - Gini is designed with proper permissioning that matches how your team operates.\n\n**2 minutes to start.** Add Gini to Slack, invite it to the channels where work happens, give it a job. It’ll ask for the tools it needs (whatever the job touches). It only reads the channels you invited it to.\n\n**Built by software engineers who’ve shipped products used by millions.**\n\nThe three of us worked together at Lilac Labs (YC S24) and all went to Berkeley. Tony (CEO) was software at Tesla and Intel. Shelden (CTO) was at Flatiron Health, Roche, and Certik (cyber security). Wilson (Chief Engineer) was at Hewlett Packard Enterprise and Vestible, a regulated securities platform.\n\nGini is the coworker we wanted at our last company, when we were the ones running every job.\n\n**Try Gini today**\n\nTry it at [ginicomputer.com](https://ginicomputer.com/?utm_source=yc&utm_medium=referral&utm_campaign=yc-launch&utm_content=try-cta).\n\nTake 30 minutes with me: [book a demo.](https://cal.com/tony-kam-ympx2p/intro?overlayCalendar=true)\n\nWould love to hear your thoughts and feedback. You can email me at [tony@ginicomputer.com](mailto:tony@ginicomputer.com)\n\n![uploaded image](/media/?type=post&id=111423&key=user_uploads/802295/f974cf1c-1978-4418-a1e6-5fdc6de258f8)\n\n", "url": "https://www.ycombinator.com/launches/Sz9-gini-ai-coworker-that-lives-in-slack", "votes": 101, "created_at": "2026-08-20T13:00:00.616Z" } ] } ]