概要
What is GitMCP?
GitMCP is a free, open‑source, remote Model Context Protocol (MCP) server that transforms any GitHub project (repositories or GitHub Pages sites) into a documentation hub. It enables AI tools like Cursor to access up‑to‑date documentation and code, eliminating code hallucinations.
How to use GitMCP?
Choose a URL format: gitmcp.io/{owner}/{repo} for a specific repository, {owner}.gitmcp.io/{repo} for a GitHub Pages site, or gitmcp.io/docs for dynamic access to any repository. Then add that URL as an MCP server in your AI assistant (Cursor, Claude Desktop, Windsurf, VSCode, Cline, Highlight AI, Augment Code, or Msty AI) using the provided configuration snippets. No downloads, installations, or signups are required.
Key features of GitMCP
- Grants AI assistants access to the latest documentation on any GitHub project
- Eliminates AI hallucinations with accurate, grounded responses
- Zero setup – runs entirely in the cloud
- Includes an embedded in‑browser chat for quick documentation browsing
- Open‑source, free to use, and does not collect personal data or store queries
Use cases of GitMCP
- Adding real‑time documentation to a Cursor or Windsurf workflow for a specific library
- Enabling an AI assistant to answer questions about any GitHub repository on the fly
- Using the embedded chat to browse a project’s docs without leaving the browser
- Self‑hosting GitMCP for full control over data and privacy
FAQ from GitMCP
Does GitMCP require any installation?
No. GitMCP runs in the cloud. You only need to add the GitMCP URL to your AI assistant’s MCP configuration.
Does GitMCP store my queries or personal data?
No. GitMCP does not collect personal information or store queries. It is designed with privacy in mind and can be self‑hosted.
Can I host GitMCP myself?
Yes. GitMCP is open‑source and you can self‑host it if you prefer.
What transport does GitMCP use?
GitMCP uses Server‑Sent Events (SSE), as shown in the VSCode configuration ("type": "sse").
What documentation sources does GitMCP support?
It supports, in order of priority: llms.txt, an AI‑optimized version of the project’s documentation, and the README.md / root.
