Review launches as open-source desktop app for local AI code review

Review, a new free desktop application for reviewing Git branches and GitHub pull requests, has launched as an open-source (MIT) project. The tool allows developers to integrate AI reviewers using their own API keys or local models, keeping all comments private on the user's machine until explicitly published.

Local AI Integration

The app supports multiple AI providers, including OpenAI, Anthropic, Gemini, and OpenRouter, as well as local models via Ollama or LM Studio. API keys are stored securely in the system keychain. A distinctive feature is the ability to split reviews across a team of models, with visibility into exactly which rules each model checked.

Rule-Based Reviewing

Unlike generic AI assistants, Review checks changes against a fixed set of rules covering bugs, security, error handling, and tests. It displays which rules were successfully checked and allows users to retry only the rules that failed or were missed. The tool can also call MCP servers to fetch context, such as ticket details, that are not present in the code itself.

Availability

Review is available for macOS, Windows, and Linux. It requires no account and is free to use. Users should note that installers are currently unsigned, requiring a manual override on first launch.