Launching today
CC Auto Review
macOS app that reviews GitHub PRs w/ your local Claude Code
13 followers
macOS app that reviews GitHub PRs w/ your local Claude Code
13 followers
A macOS app that reviews GitHub pull requests with your local Claude Code. AI findings sit beside the diff, and you accept, reject or edit each one before posting. Every PR gets a 0–100 risk score, and an optional autopilot clears routine reviews while handing risky PRs to you with nothing posted. No clone and no API key to set up: it uses your existing gh and claude. Free and open source (MIT).
Free
Launch Team / Built With
AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted
Hi Product Hunt! CC Auto Review puts a pull request's diff and Claude's findings on one screen.
How it works
It runs your installed claude -p and gh. It stores no Claude credentials and there's no account to create.
Only the PR's changed files go into a temporary directory, so you don't need a clone.
blocker and should findings must say how the problem happens and what goes wrong, or they're downgraded to a question.
The risk score starts from the diff alone. Claude can move it by up to ±20, citing file:line.
A PR can't change its own review rules: conventions come from the base branch, and Claude runs with a short list of permissions.
Try the score without installing: paste any public PR URL on the website. It runs the app's scoring code in your browser.
Limits
macOS only, and it needs Claude Code 2.1.314+ and GitHub CLI.
The autopilot runs only while the app is open, and its reviews post from your GitHub account. It runs Claude Code automatically, so check Anthropic's terms first or use an API key.
Not affiliated with Anthropic.
The screens in the gallery use sample data.
I'd love your feedback on: which PRs would you never let an AI approve? That decides what the autopilot should hand back by default.
Good luck with the launch!
I am kinda confused, what's the difference between that and having my Claude connected to GitHub through the MCP? I mean, I build locally with Claude and he even manages the branching out and PR creation, and of course he has access to my Git, so where is CC Auto Review helping it?
No offense of course, I am hoping I missed the point and I can get value from it :-)
@matan_hazan Thanks, and no offense taken! It's a fair question.
You didn't miss anything on the Claude side: CC Auto Review runs the same local Claude Code, and your setup can already create PRs and even review them if you ask. The difference is the workflow around reviewing, especially when many PRs are coming in:
You stay in control of what gets posted. Findings appear next to the diff, and you accept, reject or edit each one before anything goes to GitHub.
You can see where to focus. Every PR gets a 0–100 risk score. An optional autopilot clears routine reviews and hands risky PRs to you with nothing posted.
Reviews run with tight limits. Claude gets restricted tools and only the PR's changed files, plus any CLAUDE.md, AGENTS.md or REVIEW.md from the base branch, so a PR can't rewrite its own review rules.
If you mostly build solo and asking Claude to review your own work is enough, you may not need it. It's more useful when you're reviewing a steady stream of PRs, like teammates' PRs or PRs written by AI agents.
Happy to hear how it compares if you give it a try!