Building the future of codebase understanding
Code-Graph-RAG was born from a simple frustration: understanding large, multi-language codebases shouldn't require reading every file.
The Problem
Modern monorepos span multiple languages, frameworks, and thousands of files — often with services in Python, frontends in TypeScript, infrastructure in Go, and shared libraries in Rust or C++. As these codebases grow, they bring compounding challenges: tangled cross-service dependencies, unclear ownership boundaries, and changes that ripple unpredictably across language barriers.
Understanding which functions call what, how modules depend on each other across languages, or where a single change might break downstream consumers becomes nearly impossible to reason about manually. Teams waste hours tracing call chains through unfamiliar code in languages they don't specialize in.
Traditional tools like grep and IDE “find references” only work within a single language. They can't answer questions like “what's the call chain from the TypeScript API endpoint through the Python service to the database?” or “which functions handle authentication across all our services, regardless of language?”
The Solution
Code-Graph-RAG parses your entire codebase using Tree-sitter, builds a comprehensive knowledge graph of all code entities and their relationships, and lets you query it using natural language. It's like having a senior engineer who has read every line of your code and can instantly answer any structural question.
Beyond querying, it can surgically edit code, optimize your codebase with AI-powered suggestions, and integrate directly into your workflow through Claude Code's MCP protocol.
Open Source First
The core tool is and will always be free and open source under the MIT License. Enterprise services fund continued development.
Language Agnostic
One unified graph schema works across all 11 supported languages. Your polyglot codebase gets first-class treatment.
Community Driven
Built with community feedback. Language parsers, features, and improvements are contributed by developers worldwide.
Created by Vitali Avagyan
Vitali built Code-Graph-RAG to solve real-world codebase comprehension challenges faced by development teams working with large, multi-language projects.