Sometimes you don’t write the docs yourself – you simply want the full, up‑to‑date documentation of an open‑source library on GitHub available to your AI assistant. The community projects below turn any repository into a remote MCP server that streams documentation on demand.
First-Time Processing – If a repository hasn’t been previously indexed by Context7 or GitMCP, it may take some time (typically a few minutes) for the initial processing to complete as the system performs autoindexing and RAG preparation of the content.

​
Available Services

opengraph.githubassets

Context7

GitHub Repo starsGitHub LicenseGitHub last commit
opengraph.githubassets

GitMCP.io

GitHub Repo starsGitHub LicenseGitHub last commit

​
Context7

Context7 Context7 automatically ingests popular open‑source libraries (and many private ones too) and exposes them as remote MCP servers.
Quick Setup – Use the configuration below to get started with Context7 in your preferred environment.
Add a MCP server in ~/.cursor/mcp.json:
{
  "mcpServers": {
    "context7": {
      "command": "npx",
      "args": ["-y", "@upstash/context7-mcp@latest"]
    }
  }
}

​
Example: Using Context7 with Three.js

After installing Context7, you can prompt your AI assistant to use Three.js documentation like this: 
Can you help me create a basic Three.js scene with a rotating cube?
Use context7
The AI will then dynamically query Context7 for relevant Three.js documentation as it helps you.

​
GitMCP.io

GitMCP GitMCP lets you turn any public GitHub repository into a fully‑featured MCP server just by changing the domain from github.com → gitmcp.io (or github.io → gitmcp.io). No build step, no package installs.
Pro Tip – Simply change GitHub URLs to GitMCP URLs to instantly access documentation!
{
  "mcpServers": {
    "threejs": {
      "url": "https://gitmcp.io/mrdoob/three.js"
    }
  }
}

​
Example: Three.js Documentation

Once either MCP server is configured, the following single‑shot prompt in Cursor + Claude illustrates the improvement:
Build a Three.js scene featuring a controllable realistic person navigating a textured dynamic urban environment with realistic lighting and subtle bloom effects. Ensure keyboard controls (WASD) for movement.

​
Side‑by‑Side Result (GitMCP vs. No GitMCP)

Why it matters – by streaming the full Three.js documentation on demand, GitMCP gives the LLM precise API details (e.g. GLTFLoader, DirectionalLight.castShadow, EffectComposer) that would otherwise be unavailable or outdated, resulting in dramatically higher‑quality code generation.

​
Benefits of GitHub-Powered MCP Servers

These tools offer several advantages:
  • Zero Documentation Effort: Use existing documentation from any GitHub repository
  • Always Up-to-Date: Documentation is pulled directly from the latest repository version
  • On-Demand Access: AI assistants query only what they need when they need it
  • Broad Compatibility: Works with virtually any open-source project on GitHub
  • Simplified Integration: Easy to add to your IDE or AI assistant