Self-hosted AI code review
AI code review on every pull request, on your own servers
Register a repo and PR Bot reviews its new pull requests: a verdict, issues with severity and file, and notes on the PR as a whole. Developer replies become learning signals that shape the next review of that repo.
majorsrc/webhooks/send.ts:42
Retry loop has no upper bound; a dead endpoint retries forever.
minorsrc/webhooks/send.ts:58
Backoff delay is not jittered.
PR-level note · Test coverage: no test exercises the retry path.
Sample data · 2 issues · 1 PR-level note
Review, learn, monitor
- 01
AI code reviews
A verdict, Issues with severity and location, and PR-level notes for every review.
- 02
Watched repos
Register a repo once; PR Bot discovers its open PRs and reviews them.
- 03
A learning loop per repo
Developer replies become learning signals; approved ones shape the next review.
- 04
Lessons learned and directives
Turn a PR’s review threads into concise rules for the repo’s CLAUDE.md.
- 05
Stale review cleanup
Analyse open PRs that have gone quiet and get a recommendation for each.
- 06
Monitored PRs
Track an Azure DevOps PR against an approval threshold, with MS Teams alerts.
- 07
Reviewer activity and bypasses
Reviewer and author leaderboards, plus detection of removed negative votes.
- 08
Build monitors
Health dots for the recent runs of the Azure DevOps pipelines you watch.
- 09
Release failure analysis
Diagnose failed Playwright runs in Azure DevOps releases and post a Teams card.
- 10
Chat with your reviews
An AI chat scoped to a review, builds, monitored PRs, signals or feedback.
- 11
Feedback evaluation
Paste human review feedback and get each item sorted by what to do with it.
From open PR to a better next review
- Step 1
Watch a repo
Register a GitHub or Azure DevOps repository as a watched repo in Settings. PR Bot discovers its open pull requests.
- Step 2
Review the PR
A Temporal workflow clones the repo, fetches existing comments and runs the AI review through your MCP provider.
- Step 3
Post the result
The verdict, Issues and PR-level notes are stored and posted back to the PR as bot threads, unless it is a dry run.
- Step 4
Watch the replies
Review monitoring polls the PR for new human replies to bot threads, new commits and state changes.
Your repos, your servers, your AI back end
Run PR Bot with Docker Compose or as a macOS desktop app. Reviews go through the MCP provider you configure.