AI Review is an open-source, AI-powered code review tool that integrates with CI/CD pipelines to enhance code quality and accelerate development workflows. It supports major platforms like GitHub, GitLab, and Bitbucket, offering automated inline comments, summary reports, and AI-generated replies to improve consistency and speed up reviews while maintaining human oversight.

How It Works

The tool operates client-side within CI/CD environments, connecting to various LLM providers (OpenAI, Claude, Gemini, Ollama, Bedrock, OpenRouter, Azure OpenAI) and VCS platforms (GitHub, GitLab, Bitbucket, Azure DevOps, Gitea). It offers distinct review modes: inline comments, cross-file context analysis, and high-level summaries. A key feature is its ability to generate replies within existing review threads, fostering dynamic discussions. Configuration is flexible via YAML, JSON, or environment variables, with code data sent directly to the LLM endpoint or a local Ollama instance for privacy.

Quick Start & Requirements

Installation is via pip (pip install xai-review) or Docker. A configuration file (e.g., .ai-review.yaml) is mandatory, detailing LLM and VCS credentials. CI/CD integration examples are provided for GitHub Actions and GitLab CI/CD. Comprehensive documentation for configuration, CLI, and CI/CD is available in the ./docs directory.

Highlighted Details

• Broad LLM provider support: OpenAI, Claude, Gemini, Ollama, Bedrock, OpenRouter, Azure OpenAI.
• Extensive VCS integration: GitHub, GitLab, Bitbucket, Azure DevOps, Gitea.
• Advanced review modes: inline, context, summary, and AI replies to threads.
• Client-side execution ensures code privacy, with options for local Ollama deployment.

Maintenance & Community

Maintained by NikitaFilonov. Specific community channels, contributor details, sponsorships, or roadmap information are not detailed in the provided README.

Licensing & Compatibility

The README does not explicitly state the project's license. This requires further investigation to determine terms for commercial use and derivative works.

Limitations & Caveats

Users are responsible for securing API tokens and preventing accidental exposure of sensitive data. Reliance on external LLM providers may incur costs. The extensive configuration options may require a learning curve.