CodeGraphContext – An MCP server that converts your codebase into a graph database

CodeGraphContext- the go to solution for graph-code indexing 🎉🎉...

It's an MCP server that understands a codebase as a graph, not chunks of text. Now has grown way beyond my expectations - both technically and in adoption.

Where it is now

• v0.4.0 released
• ~3k GitHub stars, 500+ forks
• 50k+ downloads
• 75+ contributors, ~250 members community
• Used and praised by many devs building MCP tooling, agents, and IDE workflows
• Expanded to 15 different Coding languages

What it actually does

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.

That means: - Fast “who calls what”, “who inherits what”, etc 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.

• Python package→ https://pypi.org/project/codegraphcontext/
• Website + cookbook → https://codegraphcontext.vercel.app/
• GitHub Repo → https://github.com/CodeGraphContext/CodeGraphContext
• 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.

Happy to hear feedback, skepticism, comparisons, or ideas from folks building MCP servers or dev tooling.

Original post (for context):
https://www.reddit.com/r/mcp/comments/1o22gc5/i_built_codegraphcontext_an_mcp_server_that/