Context7

Context7 brings current library documentation and code examples into an AI coding assistant's context. It's for developers who want answers grounded in the libraries and versions they're actually using, rather than an assistant's older training data. It works with Claude Code, Codex, Cursor, Devin Desktop and Antigravity.

You can request documentation for a particular library version and retrieve material relevant to your coding question. Context7 searches its library index to find the right project, then supplies documentation for the task. The index includes Next.js, React, Playwright, Stripe, Supabase and Prisma. This gives an assistant source material to consult when suggesting APIs or writing code, though it doesn't guarantee that the resulting code is correct.

There are two ways to connect it to an agent: a CLI with skills that guide documentation retrieval, or an MCP server that exposes documentation tools directly. The MCP server can run locally, but documentation retrieval uses Context7's online service. It's a documentation connection for your coding assistant, rather than a runtime for running models on your hardware.

The public MCP client is MIT-licensed TypeScript software. For developers building their own AI tools, it also provides a REST API, a TypeScript SDK and tools for the Vercel AI SDK. Library entries include community contributions, and Context7 doesn't guarantee their accuracy or completeness.

A separate licensed on-premises stack includes parsing, indexing, local vector storage, a web UI and an MCP server. Deploy it in Docker or Kubernetes and configure the AI provider; local OpenAI-compatible models are supported. Self-hosting the public MCP client alone still uses the online service.