Context7: Up-to-Date Documentation for AI Code Editors
Dec 7, 2025
mcpai-toolsdocumentationupstash
Context7 is an MCP server by Upstash that delivers current, version-specific documentation directly into your AI coding assistant’s context. Instead of relying on stale LLM training data, Context7 fetches real documentation from the source.
The Problem It Solves
LLMs are trained on snapshots of documentation that become outdated. When you ask for help with a library, you might get:
- Hallucinated APIs that don’t exist
- Deprecated patterns from old versions
- Code examples that won’t compile
Context7 fixes this by injecting live documentation into the prompt.
How It Works
The pipeline: Parse (extract code snippets) → Enrich (add metadata via LLM) → Vectorize (embed for semantic search) → Rerank (score relevance) → Cache (serve from Redis).
Two MCP tools are exposed:
resolve-library-id- matches “react” to a specific documentation IDget-library-docs- fetches the actual docs
Supported Libraries
Over 33,000 libraries indexed, with community contributions for more. You can also add support for niche libraries not yet included.
Editor Integration
Works with Cursor, VS Code (Copilot), Claude Code, Windsurf, and other MCP-compatible clients.
Claude Code setup (remote):
claude mcp add --transport http context7 https://mcp.context7.com/mcp
Claude Code setup (local):
claude mcp add context7 -- npx -y @upstash/context7-mcp
Usage
Add “use context7” to your prompts:
“Create a Next.js middleware that checks for a valid JWT in cookies. use context7”
Or configure auto-invocation rules in your editor to trigger it automatically.
Pricing
Free and open source. An optional API key provides higher rate limits and private repository access.
Security Note
Since documentation is community-contributed, there’s a potential vector for “poisoned” documentation submissions. Backslash Security flagged this as a consideration when using the service.
Sources: