ai-cr 0.6.0

LLM-agnostic GitHub AI Code Review Tool with integration to GitHub actions

🤖 AI Code Review Tool

An AI-powered GitHub code review tool that uses LLMs to detect high-confidence, high-impact issues—such as security vulnerabilities, bugs, and maintainability concerns.

✨ Features

• Automatically reviews pull requests via GitHub Actions
• Focuses on critical issues (e.g., bugs, security risks, design flaws)
• Posts review results as a comment on your PR
• Can be used locally; works with both local and remote Git repositories
• Optional, fun AI-generated code awards 🏆
• Easily configurable via .ai-code-review.toml in your repository root
• Extremely fast, parallel LLM usage
• Model-agnostic (OpenAI, Anthropic, Google, local PyTorch inference, etc.)

See code review in action: example

🚀 Quickstart

1. Review Pull Requests via GitHub Actions

Create a .github/workflows/ai-code-review.yml file:

name: AI Code Review
on: { pull_request: { types: [opened, synchronize, reopened] } }
jobs:
  review:
    runs-on: ubuntu-latest
    permissions: { contents: read, pull-requests: write } # 'write' for leaving the summary comment
    steps:
    - uses: actions/checkout@v4
      with: { fetch-depth: 0 }
    - name: Set up Python
      uses: actions/setup-python@v5
      with: { python-version: "3.13" }
    - name: Install AI Code Review tool
      run: pip install ai-code-review==0.5.0
    - name: Run AI code analysis
      env:
        LLM_API_KEY: ${{ secrets.LLM_API_KEY }}
        LLM_API_TYPE: openai
        MODEL: "gpt-4.1"
        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
      run: |
        ai-code-review
        ai-code-review github-comment --token ${{ secrets.GITHUB_TOKEN }}
    - uses: actions/upload-artifact@v4
      with:
        name: ai-code-review-results
        path: |
          code-review-report.txt
          code-review-report.json

⚠️ Make sure to add LLM_API_KEY to your repository’s GitHub secrets.

💪 Done!
PRs to your repository will now receive AI code reviews automatically. ✨
See GitHub Setup Guide for more details.

2. Run Code Analysis Locally

Install and run:

# Prerequisites: Python 3.11+
pip install ai-code-review

# One-time setup using interactive wizard (saves configuration in ~/.env.ai-code-review)
ai-code-review setup

# Run review on committed changes in current branch vs main
ai-code-review

To review a remote repository:

ai-code-review remote --url https://github.com/owner/repo --branch feature-branch

🔧 Configuration

Change behavior via .ai-code-review.toml:

• Prompt templates, filtering and post-processing using Python code snippets
• Tagging, severity, and confidence settings
• Custom AI awards for developer brilliance
• Output customization

You can override the default config by placing .ai-code-review.toml in your repo root.

See default configuration here.

💻 Development Setup

Install dependencies:

make install

Format code and check style:

make black
make cs

Run tests:

pytest

🤝 Contributing

Looking for a specific feature or having trouble?
Contributions are welcome! ❤️
See CONTRIBUTING.md for details.

📝 License

Licensed under the MIT License.

© 2025 Vitalii Stepanenko