tirth8205

tirth8205 / code-review-graph

Public

Local knowledge graph for Claude Code. Builds a persistent map of your codebase so Claude reads only what matters — 6.8× fewer tokens on reviews and up to 49× on daily coding tasks.

ai-coding claude claude-code code-review graphrag
28
6
100% credibility
Found Mar 10, 2026 at 24 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A tool that builds a persistent map of codebase structure to give AI coding assistants precise, efficient context for reviews and edits.

How It Works

1
💡 Discover a code smarts booster

While chatting with your AI coding buddy about reviewing your project, you hear about a helper that maps your whole codebase for quicker, sharper insights.

2
📦 Add it to your project easily

You install the helper with one quick step, and it fits right into your coding workspace.

3
🗺️ Build your project's map

Tell it to scan your files once, and it creates a clear picture of all functions, classes, and how they connect.

4
🔄 Updates happen automatically

From now on, whenever you change code or save a commit, the map refreshes itself quietly in the background.

5
Use it with your AI buddy
🔍
Review recent changes

Get a focused look at what your edits affect, spotting risks fast.

📊
Explore connections

Search for related code pieces or see who calls what.

🖼️
View the big picture

Open an interactive map to browse your project visually.

🚀 Enjoy lightning-fast reviews

Your AI delivers spot-on advice without rereading everything, making coding reviews a breeze and saving tons of time.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 24 to 28 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is code-review-graph?

This Python tool builds a local knowledge graph of your codebase using Tree-sitter parsers across 12 languages, feeding Claude Code precise structural context like functions, calls, imports, and inheritance. It solves the token waste of AI re-scanning entire repos by delivering blast-radius analysis and relevant snippets only—benchmarked at 6.8x fewer tokens for code reviews and up to 49x for tasks. Install via pip or Claude plugin, then run CLI commands like `code-review-graph build` or slash commands for incremental updates and visualization.

Why is it gaining traction?

It stands out as a local GitHub Copilot alternative for code reviews, skipping cloud dependencies unlike Graphite code review platforms with their pricing and valuation concerns. Developers hook on auto-updates via git hooks and watch mode, semantic search, and interactive D3 graphs, plus MCP tools that Claude calls automatically for impact queries. Token savings scale massively in monorepos, making AI coding feel snappier without external runners or servers.

Who should use this?

Claude Code users reviewing PRs or fixing bugs in mid-to-large Python, TypeScript, or Go repos, especially those hitting token limits on httpx/FastAPI-scale projects. Backend teams analyzing change impacts, or solo devs wanting a local GitHub actions runner for graph-powered queries during daily tasks. Skip if you're not in the Anthropic ecosystem.

Verdict

Worth a spin for Claude Code workflows—solid benchmarks, CLI, and docs despite 23 stars and 1.0% credibility signaling early maturity. Test on your repo; low risk, high upside if token efficiency matters.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.