Crit
AI-powered bug feedback for iOS development
Capture screenshots from iOS Simulator, annotate bugs visually in the browser, and feed structured feedback directly to your AI coding agent. No chat windows, no copy-paste, no context switch.
About Crit
Crit is a command-line tool that streamlines bug reporting for iOS app development. It captures screenshots directly from the iOS Simulator, lets you annotate issues visually in the browser, and converts that feedback into structured JSON that AI coding agents can process and act on.
It works with any AI coding assistant that reads images and JSON — Claude Code, Cursor, Gemini, and more. No proprietary plugins or MCPs required.
How It Works
Capture
Run crit capture to grab a screenshot from the iOS Simulator.
Annotate
Run crit serve to open a browser UI where you click and mark problem areas.
Feed to Agent
Annotated screenshots and structured JSON are saved to .crit/ for your coding agent to pick up.
Features
Zero Vendor Lock-in
Uses open formats — images and JSON. Works with any AI agent that can read files.
Minimal Dependencies
Built on vanilla Node.js with no external npm packages.
Agent Compatible
Works with Claude Code, Cursor, Gemini, and any assistant that reads images and JSON.
Visual Annotation
Browser-based interface for clicking and marking problem areas directly on screenshots.
Who It's For
- iOS developers using AI coding agents in their workflow
- Teams that want structured, visual bug feedback without extra tooling
- Anyone who wants to close the loop between "I see a bug" and "the agent fixes it"
Support
Having trouble or have a question? We're here to help.