Open Source Startups funded by Y Combinator (YC) 2025

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.

Velum is an open-source firewall for content-level access control across documents, databases, and applications. Admins define policies, and we apply them in real time across your entire stack to control what each recipient sees. Our firewall sits between your data sources and consumers (LLMs, users, vendors), enforcing policies on every request. We understand meaning in text, documents, and databases to detect sensitive information based on your policies, then apply recipient-specific controls without blocking workflows. Create policies in natural language. Teams write rules that define what's sensitive and who should access it, and Velum enforces them consistently across all systems. When authorization changes, we can reveal tokenized data just-in-time for approved recipients, maintaining both security and utility. We integrate directly into your existing infrastructure. For AI workflows, we intercept prompts, responses, and retrievals to prevent sensitive information from leaking into or out of models. For enterprise systems like Salesforce, SAP, or Workday, we extend policy enforcement to exports, reports, and syncs. Velum is designed for scale and compliance. We provide encryption and structure-preserving tokenization for high-throughput pipelines, with multi-cloud deployment options and configurable data residency per tenant to meet your compliance requirements.

Expected Parrot helps companies simulate their customers with AI agents to explore pricing, product, marketing, communications and other scenarios at scale. Our open-source library and no-code web app let you design custom agent personas and conduct interviews and surveys with them using LLMs of your choice to generate results and reports. Everything is cached and reproducible, and you can send the same surveys to human respondents to validate results in the same interface.

AI researchers should be pushing the boundaries of what's possible with new architectures and training methods. Instead, they waste weeks configuring cloud infrastructure, debugging distributed systems, and optimizing their GPU code. We know because we lived it: While training our own models across thousands of GPUs earlier this year, we spent more time fighting our infrastructure than doing actual research. That why we're building two things. First, Elastic Cloud: a managed platform that automatically finds the cheapest GPUs across all providers, handles spot instance preemption, and cuts compute costs by up to 80%. Second, automatic kernel optimization that makes training code run faster by modeling hardware topology, often beating hand-tuned implementations. The problem is that getting high performance across different hardware is genuinely hard. NVIDIA's CUDA moat exists because writing fast kernels requires deep expertise. Most teams either accept vendor lock-in or hire expensive kernel engineers. Our goal is to break the CUDA moat. The compute bottleneck is the biggest constraint on AI progress. NVIDIA can't manufacture enough GPUs, and their monopoly keeps prices astronomical. Meanwhile, AMD, Google, and Amazon are shipping capable alternative hardware that nobody uses because the software is too hard. We're breaking that moat. If we succeed, anyone will be able to train state-of-the-art models without thinking past their PyTorch 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.

Metorial is the infra layer for AI integrations. We built the only serverless MCP runtime that can maintain state with sub-second cold starts, built-in security, and monitoring. We make it easy for developers to connect AI agents to apps and data sources at scale.

Epicenter is an ecosystem of open-source, local-first apps. We store all your work—notes, transcripts, chat histories—in a single folder of plain text and SQLite. Every tool we build shares this memory: your text editor, personal assistant, etc.. It's open, tweakable, and yours. Grep it, open it in Obsidian, host it wherever you like. The choice is yours. Our vision is to build a personal workspace where you own your data, choose your own AI models, and replace siloed apps with open, interoperable alternatives. All while preserving authenticity and being free and open source. ❤️

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.

mcp-use provides open-source dev tools and infrastructure for MCP to help dev teams quickly build and deploy custom AI agents with MCP servers. Our library has over 170,000 downloads and 7,000 stars, and is used by hundreds of engineers at big enterprises. We provide a hosted and self-hosted platform that manages auth, access control, and multi-user environments for secure MCP agents deployment at scale.

CUDA for phones & consumer hardware: Cactus is used by 3k+ developers and 100+ products to run AI locally on consumer hardware. It is open-source! Check out the repo: https://github.com/cactus-compute/cactus

stagewise builds an open-source frontend coding agent for real codebases. We enable backend devs, designers and product managers to build pretty and functional web apps in a simple and visually guided manner - no matter the used tech stack. stagewise lives right inside your browser, makes changes in your local codebase and is compatible with all kinds of frameworks and setups - allowing it to be retrofitted into any existing project. Our agent enables you to show and tell what you want to change - all while using your app on localhost and without having to switch to external tools. Making changes is as easy as selecting elements, giving a simple prompt like “make it green and put it in the top right corner”, and immediately seeing the result. We're two founders from Germany who met while working at a B2B SaaS startup. With Glenn's background in building dev tools for enterpries (Rheinmetall, Miele), as well as Julians experience in building web apps used by thousands (saasmetrix, strukto.io), we have the unique blend of experience and product sense needed to build this. After making our initial launch with the open-sourced toolbar that puts your favorite coding agent right inside your browser and going viral with it, we’re now launching a high-speed coding agent dedicated to frontend development - taking our product and the experience of our users to the next level. Our vision is to replace frontend devs completely, by gradually enabling both the backend devs, but also less technical designers and product managers to create well-designed applications. We start with developers who lack design experience, but our ultimate goal is to change how products are built.

As AI agents replace traditional software, they need a way to store, recall, and reason over contextual data. HelixDB gives them all of that in one system. We’re building the knowledge infrastructure every AI agent will depend on.

Zero empowers knowledge workers by transforming the way they manage their email. It intelligently summarizes important messages, helps complete tasks, and even enables users to chat directly with their inbox.

Morphik is open-source multimodal search for AI apps. We're used by customers ranging from space-tech teams searching across research papers to developers building brokerage agents.

Better Auth is a comprehensive authentication framework for TypeScript that lets you implement everything from simple auth flows to enterprise-grade systems directly on your own database, embedded in your backend. It’s loved by developers all over the world and endorsed by leading voices in the ecosystem. On top of the framework, Better Auth also provides an infrastructure layer to help you scale with user management and analytics, bot and fraud detection, transactional email and SMS, global session storage and more.

Open-source infrastructure for Computer-Use Agents. Sandboxes, SDKs, and benchmarks to train and evaluate AI agents that can control full desktops (macOS, Linux, Windows).

Klavis AI is Open source MCP integration platform that let AI agents use tools reliably at any scale. Our API provides hosted, secure MCP servers, eliminating auth management and client-side code. Co-founder Xiangkai Zeng co-authored the Gemini paper and pioneered the tool calling infrastructure for Gemini. https://www.klavis.ai/

Mentra is building the open source operating system for smart glasses. The smart glasses hardware is finally ready, so we're building the OS and app store to power the next computing platform.

superglue replaces brittle SQL scripts, cron jobs, and one-off connectors that power enterprise operations with clean, versioned syncs you can actually understand and trust. It rebuilds the logic behind integrations, so migrating systems, adding tools, or scaling processes no longer breaks everything. Companies running on superglue roll out AI projects faster, spend less on maintenance, and future-proof their business for continuous optimization.

Mastra is the easiest way to build a production AI application, built by the team that brought you Gatsby. We provide a complete set of AI primitives: workflows for complex operations, agents for autonomous decisions, RAG for knowledge integration, and evals for quality & accuracy. Mastra is open-source, modular, and built in Typescript for product developers. You can deploy Mastra on any cloud, and it gives you a straightforward syntax to solve the hardest problems building with AI: human-in-the-loop workflows, agent memory and tool selection, and output quality.

Confident AI allows companies of all sizes to benchmark, safeguard, and improve LLM applications, with best-in-class metrics and guardrails powered by DeepEval. Built by the creators of DeepEval (4.3k stars, >400k monthly downloads), Confident AI is able to offer battle-tested, open-source evaluation algorithms while providing the infrastructure needed for teams to stay confident their LLM systems.

We enable AI to control the browser. Browser-use is the biggest open-source web agent project. We got 40k stars in the last 3 months.

Lingo.dev helps dev teams use LLMs for app localization and content translation. Lingo.dev Compiler helps React developers localize apps without rewriting the existing code. Lingo.dev CLI helps web and mobile teams use AI for producing perfect translations for their user interfaces and docs. Lingo.dev SDK helps use LLMs for real-time translations.

We are building a fully local opensource software engineer. Software thrives on opensource tools and it's only right that this technology be built for all to use.

Unsloth helps builders create custom models better & faster using RL and custom kernels. We're developing the all in one solution to help you create highly-accurate custom models 30x faster with 90% less memory use. With over 10 million monthly model downloads and 40K GitHub stars, our mission is to make training and open-source the best it can be!

A Cursor-like AI-assisted, no-code IDE for building production-ready multi-agents. ✨ Start from a simple prompt to create fully functional agents with the Copilot 🧪 Test them in AI-simulated scenarios 🌐 Connect MCP servers and tools 📞 Interact through the Python SDK, a web widget, or a Twilio phone number ♻️ Continuously refine your agents by providing feedback to the Copilot Built on OpenAI's Agents SDK, Rowboat is the fastest way to build multi-agents!

Opslane is a developer-first Kubernetes delivery platform. Connect your repo and get previews, safe rollouts, and instant rollbacks — no YAML, no ops team required.

Ben and Jacob met in Calculus class in 10th grade, and tried founding a startup in high school — it didn’t go anywhere. During high school, Ben lead the frontend team for the World Health Organization’s COVID-19 App Team. After high school, Jacob went to Apple and Ben went to college. When there was an opening on his team, Jacob had Ben apply, and Ben left UChicago to go work at Apple. They spent about 2 years at Apple as Senior Engineers in the Conversational Engineering Department, before leaving to found Freestyle. Freestyle is directly inspired by what they saw as their failures of the tech they worked with at Apple.

Mem0 provides a memory layer for LLM applications, enabling them to remember and learn from user interactions over time.

OpenFoundry is an open source platform that helps engineers build, deploy, and scale their open source AI stack 10x faster. Ship your AI stack with just one line of code. You could think of us as an open source alternative to Hugging Face. Learn more and check out a demo of this in action in our Launch YC post: https://www.ycombinator.com/launches/KhY-openfoundry-developer-infrastructure-for-open-source-ai

Lumen let's you implement complex pricing models in under 10 minutes. Lumen will take care of webhooks, user credits, tracking usage in real-time, usage based billing, taxes and more so you can focus on your product.

We work on benchmarking and improving the long-horizon capabilities of AI Agents. We build out specialised environments to improve the long-horizon capabilities of browser and computer use agents.

Hatchet abstracts away the infrastructure for managing task queues and message brokers, so you can focus on writing your background task and workflow logic. With Hatchet, you can run slow OpenAI requests in the background with async tasks, chain complex tasks together into workflows, and set retries and timeouts to recover from failure.

Tesseral is the open source platform for managing identity and access in business software. It provides enterprise-grade capabilities, including SAML single sign-on, SCIM provisioning, role-based access control, managed API keys, and audit logs, implemented in just a few lines of code. Tesseral unifies access management for employees, customers, services, and AI agents, giving organizations the flexibility to adapt to any deployment model and the authority to enforce security policies with precision. You can learn more by reading our docs at https://tesseral.com/docs or by checking out our GitHub at https://github.com/tesseral-labs/tesseral.

nCompass is a platform for acceleration and hosting of open-source and custom AI models. We provide low-latency AI deployment without rate-limiting you. All with just one line of code.

Onyx is the open source generative AI platform connected to your company’s docs, apps, and people. We ingest and sync from all sources of information (Google Drive, Slack, GitHub, Confluence, Salesforce, etc.) to provide a centralized place for users to ask about anything. Imagine your most knowledgeable co-workers, all-rolled into one, and available 24/7! Check us out on Github: https://www.github.com/onyx-dot-app/onyx Or visit our website: https://www.onyx.app/

Tracecat helps security, IT, and production support engineers close tickets fast. We provide everything they need to automate mission-critical work — AI agents, workflow builder, and case management. Built open source on open source (Temporal, Pydantic AI, Postgres, Clickhouse). Founded in 2024, deployed by more than 200 organizations from startups to Fortune 500 and federal agencies.

Superagent makes AI products safe and compliant. We provide defense models and adversarial tests that prevent safety and compliance failures in AI agents and give companies the proof they need for enterprise deals. Our small language models, Guard, Redact, and Verify, run inside agents to stop prompt injections, data leaks, and unsafe outputs. Our tests measure prompt resistance, data protection, and factual accuracy so teams know how their system behaves under pressure.

Tusk (YC W24) is an AI agent that helps developers generate unit and integration tests with codebase context. Tusk sits in CI and provides automatic coverage whenever software engineers raise a PR. Engineering teams use Tusk to prevent critical regressions while increasing test coverage.

Mattermost provides secure collaboration for technical and operational teams that work in environments with complex nation-state level security and trust requirements. We serve technology, public sector, national defense and financial services industries with customers ranging from tech giants, to the world’s largest banks, to the U.S. Department of Defense and governmental agencies around the world. Our self-hosted and cloud offerings provide team messaging, file sharing, integrations, audio and screen share, workflow automation and project management on an open source platform deployed by the world’s most secure and mission critical organizations. We co-build the future of collaboration with over 4,000 open source project contributors who’ve provided over 30,000 code improvements towards our shared product vision, which is translated into 20 languages.

The first version of RecipeUI was a viral project at Robinhood that reached 100% adoption from all engineering and QA teams. When an API broke, anyone from Frontend to QA to PMs could debug without relying on backend engineers

Building a modern alternative to Salesforce: https://github.com/twentyhq/twenty

Khoj is an open-source, personal AI application that helps you get answers to your questions, whether they be in your own notes or online. You can use it to get answers that are grounded in reality, references included. Rather than spending an hour researching collating information from different sources, have Khoj do it for you, easily and quickly. Khoj can also generate images and understand your voice via TTS, so you can work hands-free. With our specialized agents, you can get deeper assistance on topics related to Health, Tutoring, Therapy, or something else altogether. Because we're open-source, you can always choose to self-host Khoj on your own machine for more privacy. We've previously built products for Cortana AI at Microsoft and scaled them from 0 to 20M daily active users.

ParadeDB is an open source, ACID-compliant search and analytics database. Built on Postgres for update-heavy workloads.

We are building tools for programmatic video creation. Revideo, our open-source framework, lets you create video templates in Typescript and render them via an API. It also provides a React player component to display changes in the browser in real-time.

Continue enables developers to ship faster with Continuous AI. The future of coding isn't writing more code. It's delegating the boring parts, so you can build the interesting stuff. Developers are drowning in AI-generated code. Continuous AI solves this by automating the workflows around code, not just the code itself. The role of the software engineer is shifting from prompting AI to designing processes that run automatically. With hundreds of thousands of users, Continue is loved by developers worldwide at organizations ranging from small startups to Fortune 500 companies. Continue has raised $5.1M from Heavybit, Y Combinator, and angel investors. The company was founded in 2023 and is based in San Francisco. We believe the time is now to create a future where developers are amplified, not automated.

Skyvern helps companies automate browser based workflows using LLMs and Computer Vision, fully automating manual workflows and replacing brittle or unreliable scripts. Examples include: 1. Automating materials procurement from commerce websites 2. Completing complex multi-step workflows (ie getting an insurance quote from Geico.com) 3. Automatically logging into portals and downloading invoices 4. Navigating legacy content systems to do data extraction or data entry

refine is an open-source, headless React framework for rapidly building enterprise web applications. It eliminates many repetitive tasks like ➡️ CRUD operations ➡️ State management ➡️ Routing ➡️ Authentication ➡️ Access Control ➡️ i18n and saves days of development time. So far, thousands of great projects have been built by 15,000+ monthly active refine developers. Use cases are ranging from admin panels, dashboards, and internal tools to B2B solutions and complex SaaS interfaces. refine is backed by a strong community featuring 16K GitHub stars ⭐️, 180+ contributors and 2,500 Discord members. Thanks to it's unique "Headless UI" approach, refine works with any custom design. It has also ready-made integrations with 4 most popular UI frameworks: Ant Design, Material UI, Chakra UI and Mantine. refine is totally backend agnostic and it can connect to 15+ backend services out-of-the-box, incl. custom REST & GraphQL API's. With refine, developers always have 100% control over their projects as they build, style and deploy the way they like.

Tiptap is a suite of open source content editing and real-time collaboration tools for developers building apps like Notion or Google Docs. It’s used by thousands of businesses worldwide, including LinkedIn, GitLab, Axios, and Anthropic.

At Reworkd, we're working on multimodal LLM agents that serve as the simplest way to extract web data at scale. Customers come to us with lists of 100s to 1000s of websites along with a data schema. Our agents traverse these websites, understand their structure, and generate code to extract data from them. We've been working on LLM agents since their inception and have received over 30k stars on GitHub and 1M+ users across previous agent products. If you're interested in our pilot program, shoot us an email!