CodeGraphContext indexes a repo into a repository-scoped symbol-level graph: files, functions, classes, calls, imports, inheritance - and serves precise, relationship-aware context to AI tools via MCP. v0.2.1 released ~500 GitHub stars, ~300 forks 25k+ downloads 65+ contributors That means: Fast “who calls what” queries Minimal context (no token spam) Real-time updates as code changes Graph storage stays in MBs, not GBs It’s infrastructure for code understanding, not just 'grep' search.

Ecosystem adoption: It’s now listed or used across: PulseMCP, MCPMarket, MCPHunt, Awesome MCP Servers, Glama, Skywork, Playbooks, Stacker News, and many more. A Python package→ https://pypi.org/project/codegra... Website + cookbook → https://codegraphcontext.vercel.... GitHub Repo → https://github.com/CodeGraphCont... Docs → https://codegraphcontext.github.io/ Our Discord Server → https://discord.gg/dR4QY32uYQ This isn’t a VS Code trick or a RAG wrapper - it’s meant to sit between large repositories and humans/AI systems as shared infrastructure. Still early, still evolving - but very real now. Happy to hear feedback, skepticism, comparisons, or ideas from folks building MCP servers or dev tooling.

CodeGraphContext

A powerful CLI & MCP to change code into knowledge graphs

A powerful CLI & MCP to change code into knowledge graphs