graphrag-code
A tool that indexes a TypeScript codebase into a Neo4j code knowledge graph, so Claude Code and other AI coding agents can answer architecture questions in Cypher instead of relying on embedding similarity over code chunks.
graphrag-code walks a TypeScript and TSX repository's AST, recognizes framework constructs such as NestJS controllers, modules and dependency injection, and React components and hooks, and loads the result into a Neo4j graph. An agent can then query dependency chains, endpoint inventories, component usage and dependency-injection hubs directly in Cypher.
Running its init command in a repository asks a few short questions about which packages to index and which package boundaries to enforce, then writes Claude Code slash commands, a GitHub Actions architecture-check workflow, policy configuration and a local Neo4j Docker Compose setup, and runs the first index, typically finishing in about two minutes.
Parsing is framework-aware, recognizing NestJS, React, TypeORM, GraphQL, FastAPI, Flask, Django, SQLAlchemy and generic Python classes and decorators as first-class graph nodes. Every relationship carries a confidence label with a numeric score, and the project ships a first-party MCP server with 16 tools plus installers for several coding agents.