Engineering, Product and Design Startups funded by Y Combinator (YC) in San Francisco 2026

Scale accelerates the development of AI within organizations of any size to deliver critical business insights and operational efficiency. Its data-centric infrastructure platform leverages RLHF (Reinforced Learning with Human Feedback) to help organizations build the strongest AI models that supercharge their business, with customers across industries including Meta, Microsoft, U.S. Army, DoD’s Defense Innovation Unit, Open AI, General Motors, Toyota Research Institute, Brex, Instacart and Flexport.

GitLab is the first single application for the entire DevOps lifecycle. Only GitLab enables Concurrent DevOps, unlocking organizations from the constraints of today’s toolchain. GitLab provides unmatched visibility, radical new levels of efficiency and comprehensive governance to significantly compress the time between planning a change and monitoring its effect. This makes the software lifecycle 200% faster, radically improving the speed of business. GitLab and Concurrent DevOps collapses cycle times by driving higher efficiency across all stages of the software development lifecycle. For the first time, Product, Development, QA, Security, and Operations teams can work concurrently in a single application. There’s no need to integrate and synchronize tools, or waste time waiting for handoffs. Everyone contributes to a single conversation, instead of managing multiple threads across disparate tools. And only GitLab gives teams complete visibility across the lifecycle with a single, trusted source of data to simplify troubleshooting and drive accountability. All activity is governed by consistent controls, making security and compliance first-class citizens instead of an afterthought. Built on Open Source, GitLab leverages the community contributions of thousands of developers and millions of users to continuously deliver new DevOps innovations. More than 100,000 organizations from startups to global enterprise organizations, including Ticketmaster, Jaguar Land Rover, NASDAQ, Dish Network and Comcast trust GitLab to deliver great software at new speeds.

Our mission is to make every search interaction meaningful & rewarding, through developer-friendly and enterprise-grade APIs. Algolia is the leading Search & Discovery API for websites and mobile apps. We help the most innovative companies across e-commerce, media and SaaS industries create powerful, relevant and scalable discovery experiences for their users. Unlike other solutions, our hosted platform reduces the complexities of building and scaling a fast, relevant digital experience and helps teams accelerate development time. More than 9,000 companies like Under Armour, Twitch, Periscope, Medium and Stripe rely on Algolia to manage 50+ billion search queries a month. Founded in 2012, we're backed by $184M in funding from Accel Partners, Alven Capital, Point Nine Capital and Storm Ventures. The team is headquartered in San Francisco with offices in Paris, London, New York, and Atlanta. To learn more, visit www.algolia.com. --- We’re looking for talented, passionate people to build the world’s best search technology. As an ownership-driven company, we seek team members who thrive under autonomy and diversity. We're committed to building an inclusive and diverse workplace. We care about each other and the world around us, and embrace talented people regardless of their race, age, ancestry, religion, sex, gender identity, sexual orientation, marital status, color, veteran status, disability and socioeconomic background

PagerDuty is an operations performance platform delivering visibility and actionable intelligence across the entire incident lifecycle. Their SaaS-based solution empowers over 10,000 small, mid-size and enterprise global customers such as Comcast, eHarmony, Slack and Lululemon with the insight to intelligently respond to critical disruptions for exceptional customer experience. PagerDuty was founded to deliver a new and innovative approach to increase business response and efficiency. When brand reputation depends on customer satisfaction, PagerDuty arms businesses with the insight to proactively manage incidents and events that may impact customers across their IT environment.

At Optimizely, we're on a mission to help people unlock their digital potential. We do that by reinventing how marketing and product teams work to create and optimize digital experiences across all channels. With Optimizely One, our industry-first operating system for marketers, we offer teams flexibility and choice to build their stack their way with our fully SaaS, fully decoupled, and highly composable solution. We help companies around the world orchestrate their entire content lifecycle, monetize every digital experience and experiment across all customer touchpoints – all through Optimizely One, the leading digital experience platform that powers every phase of the marketing lifecycle through a single, AI-accelerated workflow. Optimizely has nearly 1500 employees across our 21 global offices and has 700+ partners. We are proud to help more than 10,000 businesses, including H&M, PayPal, Zoom, and Toyota, enrich their customer lifetime value, increase revenue and grow their brands. At Optimizely, we live each day with a simple philosophy: large enough to serve, small enough to care. Learn more at optimizely.com.

Glue is an open source design editor for AI agents. It gives coding agents an infinite canvas to create new interfaces, iterate on existing ones, and turn those changes into production-ready code all in real time. The rise of coding agents is collapsing EPD into one role: builders. This new paradigm means design and engineering can now converge to a single workflow. Builders no longer want static mockups in one tool, and implementation in another. Glue is a collaborative workspace where builders and agents can create new components, iterate on existing interfaces, and implement as code. The same tight feedback loop that once made Figma powerful for designers is now built for builders and their armies of agents. Being open source allows us to penetrate workflows at zero cost and build trust quickly. Right now there are no design tools with AI as the primary user. Existing tools bolt on copilots as an afterthought. Soon, every company will have AI agents building their interfaces. We are building the surface for this workflow.

Ashr is a fully contained test and evals platform that improves AI agents by ensuring accuracy and quality across a wide range of user journeys through your product. Ashr generates user journeys through your agents’ tool calls, results, and questions. We don’t generate one off tests, we generate large amounts of authentic user stories through your product, and pick up on errors, inconsistencies, and failures that would otherwise take hours of manual testing or get caught by a customer. We also work to ensure agent response quality in line with your business goals by offering custom evaluation metrics. Ashr works with voice, text, automated website generation, image, file generation, and multimodal. We’re also willing to generate custom input types or make modifications for anyone interested. Currently, we integrate as an SDK, and are proactively working to make our documentation and resources as accessible and easy to use as possible.

AI coding tools have made developers 100x faster but customer-facing incidents are up 43% YoY, QA hasn't kept up. Canary is the first AI QA engineer that understands your codebase. Instead of relying on flaky DOM scraping or screenshot analysis, Canary reads your source code directly to understand developer intent and catch broken user flows before they hit production. Engineering teams go from weeks of manual testing to 90%+ coverage in days. Built by a team that built coding agents at Windsurf, Cognition and Google.

Sentrial is the monitoring platform for AI products in production. When your AI product/agent loops indefinitely, hallucinates, or frustrates users, Sentrial detects it in real-time, diagnoses the root cause, and recommends specific fixes. We provide the monitoring, evaluation, and debugging infrastructure that traditional tools weren't built for, so teams iterate based on production reality, not just offline evals.

REV1 automates the creation of manufacturing-ready 2D drawings from 3D CAD files. Mechanical engineers spend 25%+ of their time manually creating these drawings, a tedious but critical step for every manufactured part. REV1 reduces that process from hours to minutes, generating drawings with proper annotations and tolerancing so engineers can focus on design instead of documentation.

There are 2 problems: The first, CAD software is slow and the second, designing with it is a tedious process with many steps. Aurorin solves both of these problems by being a CAD software built from the ground up on a modern tech stack with AI at its heart. Building the CAD kernel lets us fully leverage AI while modernizing execution speed. Almost all existing CAD software (Solidworks, NX, Creo, Catia, Fusion 360, Onshape etc) use the same couple of underlying CAD kernels built in the 80's. It's common for these programs to take 4 hours to open a file. Aurorin uses a custom parametric and B-Rep driven CAD Kernel built to take full advantage of modern CPUs and GPUs and be AI native.

Martini is a collaborative, AI-native platform for professional filmmaking in the era of generative media. By integrating the best generative models into professional workflows, we enable world-class creatives to harness generative AI as a legitimate medium for film.

Sparkles is like Lovable for existing projects, so even Sharon from Marketing can change the button styling on the front page without going back and forth with the dev team. Someone technical connects their GitHub to Sparkles, sets up the repository and environment variables, and then anyone on the team with a @yourcompany.com email can log in on Sparkles and create pull requests without needing to know what Git is.

We’re building AI agents that do on-call support for data-heavy industries like financial services and fintech, healthcare, and insurance. On-call engineers in these industries need to inspect data to debug production issues. We monitor both data and infrastructure for issues and use AI agents to debug and suggest fixes in minutes. Data is especially sensitive in regulated industries, so we offer on-prem deployments and hardware-backed secure environments that let agents safely use production data as context while debugging. Mitch and Shipra founded Corelayer after building data infrastructure together at Goldman Sachs, where they spent many late nights and weekends debugging systems that processed 100s of billions of rows a day.

Visibl is an AI platform that turns compute into chip design capacity. As chips grow more complex, adding engineers yields diminishing returns — coordination costs rise while progress slows. The bottleneck is no longer ideas, but execution throughput. Visibl turns compute into a teammate. Our platform ingests the full design context (code, specs, scripts, tool outputs) to decode ambiguous tool feedback into actionable steps. We validate changes and package decisions for review, resulting in safer changes, shorter iteration cycles, and fewer late surprises on the path to tapeout.

Human-assisted development should be for novel ideation, not routine maintenance, bug fixes or incremental updates. Approxima is bringing you that future with agents that can scope work, code, validate their fixes end-to-end in sandbox environments and merge to production. Book a demo: https://calendly.com/ashish-selvaraj/approxima

We want software to improve itself. Sonarly connects to your monitoring stack - tools like Sentry, Datadog, and Grafana - and triages every alert to remove noise and duplicates. It then investigates logs, traces, metrics, and code to build a living map of your production system and find root cause. The result is either a ready-to-merge fix PR or a recommendation to update your alerting rules so the noise never comes back.

21st.dev is the largest React component registry for AI applications, used by 1.4M developers (200K MAU). We provide the building blocks to create AI agents and agent-powered apps: UI components, agent templates, and an SDK with built-in UI, tools, memory, observability, and managed infrastructure.

Insurers struggle to price autonomous systems because the historical data simply doesn't exist. Your car insurance draws from over 30 billion claims records, but autonomous trucks and robots have nearly zero. Valgo is the risk quantification platform that closes that gap. We are building probabilistic models of routes, tasks, and environments from the bottom up. We output a simulated loss estimate that insurers need to price coverage. Valgo is the foundational layer that gives insurance providers confidence to properly price autonomy risk. Our team combines expertise in risk estimation for autonomy with deep experience in the insurance industry. Sydney and Robert are Stanford PhDs who wrote the textbook on validating safety-critical systems and teach the course at Stanford. Robert spent 7 years at MIT Lincoln Laboratory on the core team that designed and validated an FAA-certified aircraft collision avoidance system, now a worldwide standard. Jon is a Sloan Fellow from the Stanford GSB who spent over 12 years in insurance leadership and led over $5 billion in M&A as head of corporate development for one of the largest insurers in Asia-Pacific.

AtlasGrid automatically builds a visual graph of your mobile app UI and displays analytics directly on your screens. Additionally, our fleet of AI-powered agents will continuously analyze thousands of mobile apps, gathering design insights and identifying growth opportunities + actionable product improvements.

Inspector helps front-end engineers ship faster with AI by automating the most frustrating part of the job — gathering context between the browser and the codebase. Inspector is an AI IDE for front-end that connects your browser and codebase. Inspector lets you select elements visually and prompt changes, giving coding agents access to the same runtime context developers have: visuals, elements, and logs. With Inspector, you’re editing right in the front-end, not buried under tabs and screenshots.

Lemma enables AI agents to continuously improve from user feedback and production outcomes. Instead of relying on manual prompt tuning and reactive debugging, Lemma detects when an agent’s performance drifts, pinpoints the exact step causing failures, generates optimized prompt candidates, and automatically delivers improvements through our API or by opening a pull request in your codebase. Built by ex-AI engineers who felt this pain firsthand, Lemma turns static agents into self-improving systems, helping teams ship reliable AI products faster with compounding performance gains over time. Deploy 1x. Learn forever.

Rovr streamlines engineering operations by automating intake, triage, routing, and cross-team coordination. It replaces internal forms, manual Slack triage, and scattered docs with an AI front desk that understands requests, assigns owners, updates context, and posts status changes directly into Slack. Rovr’s built-in ops CRM gives teams a single system of record for eng requests — from tooling access to platform incidents to dev-infra workflows — reducing operational drag across the entire org.

Semiotic builds AI design agents. We’re starting with agents that build landing pages for fast‑growing startups. Existing AI website builders didn't make creating websites easier—they made creating ugly websites easier. So, companies choose between blue-purple gradient slop or agencies with 3-month waitlists and $100k minimums. You don't have to choose anymore. Start a project: semiotic.com

Compyle is a less autonomous coding agent that asks before it acts. We're like Cursor or Claude Code - but you won't need to rewrite your codebase every time you use it for more than a small task. We built Compyle because we were frustrated with the direction other coding agents are headed. They're all racing to be more autonomous, running for hours without human input. This doesn't work for real production code. When agents run autonomously, they make countless small decisions that pile up in your codebase. Engineers lose ownership - they don't know what was implemented, can't effectively debug or maintain it, and struggle to build on top of it. The code becomes a black box, and they're forced to throw it away. Compyle keeps engineers in the driver's seat.

Understand user patterns and find silent failures in your AI agents. The Context Company helps teams analyze AI agent conversations to surface user behavior patterns, performance trends, and moments where agents struggle - so teams know what’s working, what’s not, and what to improve next. We turn raw conversations into structured signals: topic clustering, user feedback analysis, custom pattern tracking, and alerts. All alongside traditional observability like traces, tool calls, latency, and costs. Get started in less than 10 lines of code.

Sourcebot is an open source code understanding platform for massive codebases. We’re used by thousands of engineers in some of the largest companies in the world, including NVIDIA, Red Hat, Wikimedia, and Arista Networks. Understanding code, not writing it, is the primary bottleneck for large engineering teams. For developers, this means onboarding onto complex codebases faster. For AI agents, it means getting the necessary code context to minimize hallucinations and maximize cohesion within the wider codebase. Sourcebot solves this by giving developers and AI agents the ability to regex search across millions of lines of code instantly, as well as ask questions across thousands of repos using any flagship reasoning model. Being open source and on-prem, we can get deployed in enterprises within minutes before other tools even get approval. We met a decade ago at McGill University and have been building together full time for over 2 years. We’ve personally felt the pain of understanding large codebases while working at Microsoft, EA, Ubisoft, Google, and Meta.

Rivet is a visual AI editor for product designers to make UI changes. Designers ship the changes they make directly to their codebase. The most painful problem for designers today is that their vision doesn't match code. Rivet replaces slow collaboration between designers and engineers by letting designers just do it themselves. Rivet's editor helps turn product designers into design engineers. Design-led companies in YC are Rivet customers. Learn more at rivet.design

Our agents run hardware engineering tasks that normally take hours or days in just a few minutes, allowing engineers to process 100x more data and information. Our platform allows hardware engineers to orchestrate agents for tasks such as analyzing failures across entire test campaigns, pulling the right data and KPIs from their databases, checking how conditions (like temperature or load) affect performance and reliability, and then document findings in reports and dashboards.

We derive training data from the world's brightest minds, enabling LLMs to achieve state-of-the-art model performance.

Semble AI builds AI agents that design building-systems (fire, HVAC, plumbing, security). Construction companies use Semble to cut down weeks on commercial and residential projects. The $600B building-system integration industry still runs on spreadsheets and manual drawings, even as building codes rapidly evolve and become more localized. It’s a massive blind spot for the tech world. Jonathan and Ethan met studying computer science at Stanford. Jonathan worked at Convergint, one of the world's largest building system integrators, specifically for the purpose of learning how this industry works. Ethan is obsessed for a different reason: His house burned down in a wildfire in 2017. They are determined to make Semble the industry standard for construction design.

Stellon Labs is an AI research lab that is building tiny frontier AI models that can run on edge devices like smartphones, wearables and embedded systems. Our first open source model, KittenTTS, is a super-tiny text-to-speech model, that got 8K Github stars and 45K model downloads within 2 weeks of launching.

Panels is an audio data platform that delivers high-quality speech datasets from vetted, diverse contributors to power training and evaluation of foundational voice models.

We’re launching a small hardware device: a button that you clip to your shirt. You press it to talk to AI, and it responds instantly.

Lark lets teams write end-to-end tests in plain English. Validate dashboards, APIs, SDKs, and full product flows without maintaining fragile test code. Run continuously on feature branches, in CI, or in production.

Datafruit helps enterprise delivery teams understand their customers deeply and deliver more profitably. We build AI agents capture every conversation, structure scope from day one, and use past projects to make every engagement sharper.

Fluidize accelerates R&D for scientists and engineers by using AI to build and run simulations and experiments. Our platform speeds up development pipelines by automating setup, execution, validation, and scaling, allowing engineers to focus on their science.

Omnara brings Claude Code and Codex to your phone. Start a coding session on your desktop, then seamlessly hand off to mobile and keep building whether you're commuting, traveling, or away from your desk. Code in local or remote environments with full context preservation across devices. Plus, use voice to code hands-free by having natural conversations with Omnara to build, debug, and ship from anywhere.

SigmanticAI is building an AI-native hardware development assistant that automates the entire RTL design flow, from natural language to synthesizable HDL and testbenches, inside a seamless VSCode fork. Powered by fine-tuned Verilog LLMs, reinforcement learning, and real compiler feedback, SigmanticAI iteratively refines code until it compiles and passes synthesis, no matter how complex the design. It’s like the Cursor for HDL design. In addition to code generation, SigmanticAI can generate token-level annotations and onboarding documents to accelerate ramp-up for new engineers. With support for both cloud and on-prem deployment, it integrates with existing EDA tools to dramatically reduce debug time, improve collaboration, and accelerate hardware design across teams

Embedder is a coding agent that can write, test, and debug firmware.

Check out the project: https://github.com/cactus-compute/cactus

BootLoop is building the first AI agent to write and test firmware on real hardware. Existing AI coding assistants fail completely with embedded systems, but our agent ingests datasheets, schematics, and other design files to write firmware in your style and automatically test it on your hardware.

Most coding agents are one long conversation that compacts and loses context. Magnitude is a multi-agent system that stays sharp over long horizon tasks. Install: npm i -g @magnitudedev/magnitude

AI that optimizes AI

Idler builds reinforcement learning environments that teach AI models to code at expert human levels. We create training environments based on real-world coding scenarios that prepare models for the complex challenges they'll face in production.

Proliferate gives coding agents production-grade infrastructure- sandboxed environments, event triggers, and verification workflows- so they can ship code end-to-end, not just write it.

Cyberdesk is the first self learning computer use agent for developers to automate desktop workflows. For example, one of our customers is building an AI agent for hospital patient intake, and they’re using us to automate an entire team of people manually clicking and typing into a desktop app to book appointments. We launched mid July, and our agent is already operating in production environments, creating appointments in electronic health record apps and processing files in legacy ERPs. We’ve built the first computer use agent that memorizes tasks and repeats them quickly while still adapting to unexpected changes on the screen. A good analogy is to how humans use computers: once we get used to an app, we use it much faster, while still adapting to random popups and crashes. This new paradigm for computer use will be needed by every company automating human workflows on computers. Mahmoud previously built robotic process automation solutions impacting 20K+ employees at a Fortune 100 company. He also built AI & ML solutions at Baylor College of Medicine Neuroscience & Texas Children's Hospital + co-published a paper at the Rice Networks Group. Alan is ex-NASA and previously dropped out to be the first founding engineer at a PearX startup. He is a second time founder, having previously built and scaled a consumer AI app to thousands of paying users. Computer-use agents will automate the $6T dollars spent on enterprise knowledge work and Cyberdesk will be the platform to power this. We’d love to talk with you! https://cal.com/mahmoud-al-madi-klrs5s/30min