Argus

Open-source AI code review that learns from your team

3 followers

Open-source AI code review that learns from your team

3 followers

Argus reviews GitHub pull requests with specialist agents for bugs, security, architecture, and regressions. It adjusts review depth to each change and asks every finding to explain a concrete failure and a fix. Team feedback and past fixes become memory for future reviews. See which checks ran and what each review cost. Open source under AGPL-3.0, self-hostable, and bring your own LLM key.

Free

Launch tags:Developer Tools•Artificial Intelligence•GitHub

Launch Team / Built With

Render Sandboxes: a computer for every agentGive every agent its own machine. Starts in under a second

Promoted

Maker

📌

Hi Product Hunt! I built Argus to make code reviews useful across more than one pull request. A team's past fixes and review feedback should help catch the next bug. Argus adjusts review depth to the change, uses specialist reviewers, and asks findings to include a concrete failure scenario and a fix. It learns from reactions and replies, and shows the checks and token costs behind each review. You can self-host it and bring your own LLM key. The code is open source under AGPL-3.0. A special thank you to the qBraid team! I wrote a walkthrough of how Argus's memory works, including the diagrams and dashboard images in this gallery: https://x.com/belazyAF/status/21... I'd love to hear what your current reviewer keeps getting wrong, and what you wish it remembered about your codebase.

Report

2mo ago

Reviews