# code-review-graph

> Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters — 8.2x average context reduction in the current review benchmark and up to 49x in documented large-repo workflows.

code-review-graph (CRG) parses your repository with Tree-sitter into a structural graph (functions, classes, imports, call chains, communities), stores it locally in `.code-review-graph/graph.db`, and exposes it to AI coding tools via the Model Context Protocol (MCP) and a CLI. It works with Codex, Claude Code, Cursor, Windsurf, Zed, Continue, OpenCode, Antigravity, Gemini CLI, Qwen, Kiro, Qoder, GitHub Copilot, and GitHub Copilot CLI.

The current release is v2.3.4: 30 MCP tools, 5 MCP prompts, estimated context savings, compact architecture output, bounded large-repo analysis, Windows/FastMCP stability, and Rust test detection fixes. MIT licensed.

## Pages

- [Home](/): Product overview, how it works, benchmarks, install, supported AI coding tools, privacy, limitations.

## Optional

- [GitHub repository](https://github.com/tirth8205/code-review-graph): Source code, README, full documentation, and releases.
- [Discord community](https://discord.com/invite/3p58KXqGFN): Discussion and support.