Context7

Context7, built by Upstash, is a documentation platform and Model Context Protocol (MCP) server that feeds large language models and AI coding assistants with current, version-specific library documentation and code examples. The core problem it addresses is that LLMs are often trained on outdated snapshots of library docs, leading to hallucinated APIs or deprecated code patterns; Context7 pulls fresh reference material at query time instead. Two tools sit at the center of the workflow: resolve-library-id, which maps a library name to a Context7-compatible identifier, and query-docs, which retrieves the matching documentation for that identifier. Developers can install the MCP server directly into clients such as Cursor, Claude Code, and Claude Desktop, or browse and search documentation through the web interface, which also includes a 'Try Live' exploration mode and a ranking of popular libraries. Anyone can submit new documentation sources to expand coverage. A free API key raises rate limits, and the service also works without one for lighter use, with paid and enterprise tiers available for teams needing higher volume. Context7 is aimed at developers working across many APIs and frameworks who want their AI tools to reference accurate, current information rather than relying on a model's static training data.