GitNexus: Zero-Server Code Intelligence Engine with Built-in MCP and Graph RAG
GitNexus is an open-source code intelligence engine that runs entirely client-side, building knowledge graphs from codebases with a built-in Graph RAG Agent. The project is trending on GitHub with over 13,000 stars.
Drop in a GitHub repo or ZIP file, and GitNexus creates an interactive knowledge graph through a multi-phase indexing pipeline: structure mapping, Tree-sitter AST parsing, import/call resolution, functional community clustering, execution flow tracing, and search indexing. All processing happens in the browser — no server, no vendor lock-in, no privacy concerns.
The CLI can index a repository and run an MCP server that gives AI agents deep codebase awareness. A single command indexes the codebase, installs agent skills, registers Claude Code hooks, and creates context files. Any MCP-compatible agent — Cursor, Claude Code, Windsurf — can connect immediately.
GitNexus precomputes structure at index time, so tools return complete context in one call rather than requiring multiple queries. This makes it particularly valuable for AI agents that need structured understanding of dependencies, call chains, and execution flows.
GitHub: https://github.com/abhigyanpatwari/GitNexus
← Back to all articles
Drop in a GitHub repo or ZIP file, and GitNexus creates an interactive knowledge graph through a multi-phase indexing pipeline: structure mapping, Tree-sitter AST parsing, import/call resolution, functional community clustering, execution flow tracing, and search indexing. All processing happens in the browser — no server, no vendor lock-in, no privacy concerns.
The CLI can index a repository and run an MCP server that gives AI agents deep codebase awareness. A single command indexes the codebase, installs agent skills, registers Claude Code hooks, and creates context files. Any MCP-compatible agent — Cursor, Claude Code, Windsurf — can connect immediately.
GitNexus precomputes structure at index time, so tools return complete context in one call rather than requiring multiple queries. This makes it particularly valuable for AI agents that need structured understanding of dependencies, call chains, and execution flows.
GitHub: https://github.com/abhigyanpatwari/GitNexus