Diffray – Open-source multi-agent code review CLI (github.com)

Diffray has launched an open-source CLI tool that utilizes multiple AI agents to conduct thorough code reviews on changes in software projects. This innovative tool targets various aspects of coding—from bugs and security vulnerabilities to performance issues and code style, allowing teams to leverage specialized agents like "bug-hunter" and "security-scan." Unlike the cloud-based diffray.ai service, which automatically learns from team feedback and generates rules, the CLI version demands manual configuration, offering developers greater control over their code review process. The significance of Diffray lies in its ability to enhance code quality by combining different reviews concurrently, ensuring that only verified issues are reported after deduplication and validation. This approach not only streamlines the review process but also reduces false positives, which can lead to greater efficiency in software development. By integrating with existing AI tools like Claude Code and Cursor, Diffray becomes an adaptable solution, allowing developers to tailor their code review environment to meet project-specific needs and standards while maintaining compatibility with CI/CD pipelines through customizable JSON output options.