Argus
AI code review that actually understands your codebase
Start new thread
Meru

Argus - AI code review that actually understands your codebase

by•
Your coding agent shouldn't grade its own homework. Argus is a local-first, modular AI code review platform — one binary, six tools, zero lock-in. It combines structural analysis, semantic search, git history intelligence, and LLM-powered reviews to catch what your copilot misses.

Add a comment

Replies

Best
Meru
Maker
📌
Hey PH — I'm the creator. Got tired of AI agents writing code and then reviewing their own work. That's not review, that's grade inflation. Built Argus because I wanted an independent eye that actually understands my codebase — not just the diff, but the structure, the history, what's around it. It's local-first (code never leaves your machine), works with Gemini (free), OpenAI, or Anthropic, and plugs into Cursor/Windsurf/Claude Code via MCP. Six tools in one binary: review, map, search, history, diff, describe. v0.4.0 just shipped with feedback learning. Curious what would make this useful for your workflow. https://github.com/Meru143/argus