Open-source AI agent that automatically reviews pull requests and offers /describe, /review, /improve and /ask commands on GitHub, GitLab, Bitbucket and Azure DevOps.
cirolini/genai-code-review
A GitHub Action that reviews pull request diffs with an LLM and posts a small set of prioritized inline comments.
- Stars
- 372
- Forks
- 70
- Open issues
- 0
- Last commit
- 2 months ago
What is it?
GenAI Code Review runs as a GitHub Action on pull requests. It sends the diff to an LLM (OpenAI, Anthropic, Google Gemini, or any OpenAI-compatible server) and posts inline comments on the PR. It needs only one workflow file and no checkout step, since it reads the pull request over the API.
The project is built around a comment budget. By default it posts the top 5 findings ranked by severity and confidence, and its summary states how many findings were held back and lists them. Findings already raised on earlier pushes are filtered out before ranking, and lockfiles, generated code and similar files are skipped by default. The summary also lists files that were not reviewed, such as oversized or truncated ones.
The action supports an optional panel mode that runs two providers over the same diff and merges their findings, marked as experimental. Run logs record tokens, estimated cost, latency and posted versus suppressed findings. A separate read-only command reports how reviewers responded to the comments.
Key features
- 01Supports OpenAI, Anthropic, Google Gemini and OpenAI-compatible endpoints such as Ollama or vLLM.
- 02Posts the top few findings as one grouped inline review, with a sticky summary comment.
- 03Reports how many findings were suppressed and what files were not reviewed.
- 04Skips lockfiles, vendored and generated files by configurable glob patterns.
- 05Writes a JSON run log with tokens, cost, latency and findings.
- 06Treats diff content as untrusted and validates model output against a schema.
Use cases
- Adding an automated first-pass review to every pull request in a team repository.
- Running reviews against a self-hosted model so code never leaves your infrastructure.
- Keeping review noise low so human reviewers still read the bot's comments.
- Tracking the cost and latency of LLM reviews across pushes.
In collections
Topics
FAQ
- What is genai-code-review?
- A GitHub Action that reviews pull request diffs with an LLM and posts a small set of prioritized inline comments. GenAI Code Review runs as a GitHub Action on pull requests. It sends the diff to an LLM (OpenAI, Anthropic, Google Gemini, or any OpenAI-compatible server) and posts inline comments on the PR. It needs only one workflow file and no checkout step, since it reads the pull request over the API.
- Is genai-code-review free to use? What license does it have?
- genai-code-review is open source under the MIT license. Check the license terms before using it in commercial products.
- Is genai-code-review still maintained?
- Updated within 6 months. The last commit to genai-code-review was on Sep 21, 2026.
- What can I use genai-code-review for?
- Adding an automated first-pass review to every pull request in a team repository. Running reviews against a self-hosted model so code never leaves your infrastructure. Keeping review noise low so human reviewers still read the bot's comments. Tracking the cost and latency of LLM reviews across pushes.
- Where can I get genai-code-review?
- The source code and installation instructions are on GitHub: https://github.com/cirolini/genai-code-review
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A GitHub Action that posts AI-generated code review comments on pull requests, triggered by a /gemini-review comment, using Google's Gemini API.
A GitHub Action that runs Claude Code on PRs and issues to answer questions, review changes, and implement code.
A GitHub code review bot that uses ChatGPT-style LLMs to comment on pull request diffs, run as a GitHub App, Action, or self-hosted Probot app.
Danger JS runs after your CI to automate code review conventions, such as enforcing changelogs or PR description rules, in pull requests.
A CLI and GitHub Action that runs after AI code review to score a pull request's merge risk and list the files and endpoints it reaches.