Code Graph Rag

This open source tool builds a knowledge graph of a codebase and lets you query and edit it in plain English. It parses source files into syntax trees across 14 languages, stores functions, classes, modules, imports and call relationships in a Memgraph database, and uses Qdrant for semantic search. An AI layer translates natural language questions into Cypher graph queries, so a question such as which functions handle authentication returns the exact files and lines. It is aimed at developers and teams working in large or unfamiliar repositories. Beyond querying, it offers surgical AI code editing that targets functions by syntax tree with visual diff previews, AI assisted optimization with an interactive approval step, structural search and replace, data flow tracing that follows values from sources to input and output sinks, and dead code detection with a failure mode suited to CI. It runs as an MCP server with 15 tools for indexing, querying, editing and searching from Claude Code, and it includes a Python SDK and graph export. Installation is a Python package plus a daemon command that starts the databases. The Community edition is free and open source, with every core feature and support through GitHub. Professional and Enterprise tiers are quote based through a contact form. Professional adds a cloud hosted graph database and agent, custom language support, priority email support, SLA backed response times and onboarding. Enterprise adds custom graph relationships for compliance mapping, data lineage and ownership, on premises and air gapped deployment, custom development, training workshops and a dedicated account manager.