You pay your model provider directly. Argus takes no cut.
Questions
What does it cost?
Argus itself is free and open source. The only cost is your LLM provider — you connect your own key and pay them directly, so you see and control the spend. Self-host it and there is nothing else to pay.
Does it work with monorepos?
Yes. Argus reviews the diff of each PR regardless of repo structure. It triages files individually, so large PRs in monorepos still get fast, focused reviews.
What data does Argus store?
PR metadata (title, author, branch), the diff, review comments, and optionally past patterns for memory. We never store your full source code. Diffs are processed in-memory and discarded after review.
Can I use my own LLM API key?
Yes, and it is the intended setup. Configure any OpenAI-compatible provider — OpenRouter, OpenAI, Anthropic, Azure, Vercel AI Gateway and others — per pipeline stage from the Settings page.
What if Argus flags something incorrectly?
Dismiss it. Each dismissal is stored as a memory with the reason and the kind of change it came from — repeated dismissed patterns stop being posted (security findings are never suppressed). Every comment explains its reasoning, so you can disagree and move on.
Are any features held back?
No. There is no paid tier and nothing is gated. Deep review, simulation, memory and diagrams are per-repo settings you turn on, not upsells.
Can I run it on my own infrastructure?
Yes. Argus is open source and self-hostable — set SELF_HOSTED=true, point it at your own Postgres and GitHub App, and it runs entirely on your infrastructure with your keys.