Code Graph RAG

Code Graph RAG is an AI-powered codebase analysis tool that builds knowledge graphs for multi-language codebases.

Code-Graph-RAG is an AI-powered codebase analysis tool that builds comprehensive knowledge graphs from multi-language codebases using Tree-sitter parsing. It stores functions, classes, imports, and call relationships in Memgraph. Users can query, edit, and optimize code using natural language, translating user intent into precise Cypher graph queries. It supports 14 languages including Python, TypeScript, JavaScript, Rust, Java, and C++. It operates as an MCP server providing 15 tools for integration with Claude Code.

Code Graph RAG at a glance

Best for: Software engineer
Runs on: self-hosted, open source, API
Pricing: Free and open source community edition; enterprise support, cloud-hosted, and on-premise deployment available upon contact.
Released: June 2025
Category: AI coding tool
Website: code-graph-rag.com

Key features

Tree-sitter based AST parsing across 14 languages
Knowledge graph storage in Memgraph
Natural language queries using Cypher
AI code editing with visual diff previews
Structural search and replace via ast-grep
Data-flow tracing with FLOWS_TO taint edges
Dead code detection and reachability analysis
MCP integration with Claude Code

Use it for

Querying codebases with natural language
Performing AST-based code editing and refactoring
Tracing data-flow and identifying dead code
Integrating codebase intelligence with Claude Code

Companions

Tools that articles about Code Graph RAG recommend using with it, each with its job.

Claude Code — Installs as a skill so the agent queries the graph before grepping.
Codex — The installer wires the same graph into Codex.
Cursor — Auto-detected and configured alongside Claude Code.
OpenCode — Supported for mixed-agent setups.
Obsidian — Open the output as a vault to see how the code clusters.

Alternatives to Code Graph RAG

Axiomcodegraph — Resolves calls from types rather than syntax for higher accuracy.
Graphify — It replaces grepping with traceable graph paths, reducing token usage significantly.
jevgrep — Reduces coding agent costs by providing targeted, relevant repository context.
CodeGraphContext — Turns codebases into interactive knowledge graphs for AI assistants.
My RAG — Compare Python and Java implementations sharing identical data.
Codebase Memory MCP — High-speed indexing and persistent knowledge graphs for AI coding agents.
CodeSight — Reduces AI token usage significantly and maps projects instantly.
Rubi — It cuts AI token waste by keeping sessions short and mapping project knowledge.
llmware — Build private, local RAG pipelines with 300+ prepackaged models and integrated document parsing.
Codebase Guide — It generates accurate documentation with diagrams directly from the source code.
Code Review Graph — Provides a median 63x context reduction while keeping all code processing local.
Graft — Reduces tool calls by 46% and lowers latency by 60% for coding agents.

All Code Graph RAG alternatives