Report Abuse

Basic Information

GitMCP is an open source remote Model Context Protocol (MCP) server that turns any public GitHub repository or GitHub Pages site into a live documentation and code hub for AI assistants. It exposes repository documentation and code via MCP endpoints so tools like Cursor, Claude, VSCode and other MCP‚Äëcapable assistants can fetch up‚Äëto‚Äëdate docs, search code, and retrieve linked URL content. The project supports both repository‚Äëspecific endpoints and a dynamic generic endpoint that lets an assistant choose repositories on the fly. GitMCP is cloud‚Äëhosted by default, free to use, and can be self‚Äëhosted. It prioritizes llms.txt and an AI‚Äëoptimized documentation view, runs without authentication, and is designed to reduce LLM hallucinations by providing grounded, sourceable context to downstream agents.

Links

Categorization

App Details

Features
GitMCP provides up‚Äëto‚Äëdate documentation access, smart search, and code search as MCP tools so assistants request only the necessary context. Built features include repository‚Äëspecific and generic endpoints, embedded in‚Äëbrowser chat for repo docs, badge support to add to READMEs, and automatic view counting for tool calls. Tools exposed to assistants include fetch_documentation, search_documentation, fetch_url_content and search_code, with dynamic tool naming for the generic endpoint. Integrations and configuration examples are provided for many clients including Cursor, Claude Desktop, Windsurf, VSCode, Cline, Highlight AI, Augment Code and Msty AI. The project emphasizes zero setup for cloud usage, privacy by not storing queries or personal data, and is available under the Apache 2.0 license so it can be self‚Äëhosted and extended.
Use Cases
GitMCP helps AI assistants generate accurate, grounded responses by supplying the latest repository docs and code on demand, which substantially reduces code hallucinations and improves correctness of API usage and examples. It lets assistants search specific project documentation or code instead of guessing, enabling reliable answers even for niche or rapidly changing libraries. For teams and developers it provides an easy MCP configuration pattern and examples for popular tools, a badge to surface docs in READMEs, and local development scripts for testing. Privacy and respectful scraping are built in: public content is fetched only on request, queries are not stored, and robots.txt is respected. The service is free, open source, and suitable for both cloud use and self‚Äëhosting.

Please fill the required fields*