AI Code Review Assistant
9.9k 2026-04-14
tirth8205/code-review-graph
Optimizes AI code reviews by building a local knowledge graph of your codebase, drastically reducing token usage and providing precise context.
Core Features
Builds a persistent, structural knowledge graph of the codebase using Tree-sitter.
Tracks incremental code changes automatically.
Provides precise, context-aware information to AI assistants via Model Context Protocol (MCP).
Performs "blast-radius" analysis to identify minimal relevant files for review.
Significantly reduces AI token consumption for code reviews and daily coding tasks.
Auto-detects and configures integration with various AI coding platforms.
Quick Start
pip install code-review-graphDetailed Introduction
AI coding tools often re-read entire codebases for every task, leading to high token costs and inefficient reviews. code-review-graph addresses this by creating a local, structural knowledge graph of your project using Tree-sitter. It intelligently tracks changes and, through the Model Context Protocol (MCP), provides AI assistants with only the most relevant context. This approach drastically cuts token usage, enabling faster, more focused, and cost-effective AI-powered code reviews and development tasks across multiple platforms.