Crit

Crit, built by staff engineer Tomasz Tomczyk after growing frustrated reviewing terminal diffs from AI coding agents, is a local-first, browser-based review tool for examining AI agent output — letting you click any specific line and leave feedback the agent can read and act on directly. It supports four review modes (plans/markdown documents, code diffs across branches, live running applications, and static HTML artifacts), keeps comments attached across multiple iteration rounds so you can see exactly what changed each cycle, works with any agent that can read files and execute commands (with first-class support for Claude Code, Cursor, and GitHub Copilot), and runs entirely on localhost with zero telemetry by default. Completely free and open source — a single binary CLI with no account requirement, optional self-hosting available. Good fit for a developer tired of scrolling raw terminal diffs from AI agents who wants a proper visual review interface with persistent, line-level feedback instead.