Welcome to the ChatOSS docs. ChatOSS is a local-first desktop AI workspace — a native app where AI agents do real work alongside you: they write code in your repositories, plan and track tasks on an integrated Kanban board, and run in a suite of apps you can extend. Everything runs on your machine, and it's more than chat: it's a desktop "operating system for AI work" where each app in the dock is pluggable.
Three ways to power it
- Local Ollama models — run entirely on your machine, free forever. No account, no credit card, no limits.
- Your own API keys — bring any OpenAI-compatible provider (Anthropic, OpenAI, local servers like LM Studio or llama.cpp, and more) in Settings → API Keys. Requests route straight to that provider with your key.
- ChatOSS cloud models — a paid plan unlocks the cloud catalog. Browse the models and see pricing. Cancel anytime; local use is always free.
Download and install
ChatOSS installs on macOS, Linux, and Windows — grab the installer on the download page, or see all releases on GitHub. Local Ollama models need the Ollama daemon running on localhost:11434 — ChatOSS works with it out of the box.
What's in these docs
- For AI agents — the AI agent guide is the docs landing page: copy the complete manual in one click and paste it into any AI agent conversation.
- Getting started — what the built-in apps do.
- Build an app — everything about .aip apps: the manifest, styling, the runtime bridge, capabilities and permissions, and design guidance.
- API reference — every
window.chatossAPI, one page per capability: from AI chat and Files to Terminal, Kanban boards, and the Agent Engine.
If you're building an app with an AI agent (inside ChatOSS's Code app or anywhere else), paste it the manual from the AI agent guide — the Build an app and API reference sections together are the complete, current contract.