Enables an AI to see and operate an Expo / React Native app without an emulator, phone, or browser using real screenshots, taps, typing, scrolling, and automated bug hunting, with optional Android and iOS device control.
Enables AI assistants to capture and analyze screenshots using Claude Vision API, providing AI-powered insights about desktop interface content, UI elements, and visual layouts.
Exposes two tools over stdio or HTTP that let AI agents read display geometry and capture full-screen, per-display, or region screenshots in PNG or JPEG. Runs as a VS Code extension or a standalone MCP server usable by Claude Code, Cursor, Cline and other clients.
Enables Claude to capture screenshots, watch your screen in real-time, read text via OCR, and analyze video files, all running locally as an MCP server.
Lets MCP-capable AI agents control an Expo / React Native app running live in Expo Go on a real phone, reading a compact token-cheap outline of the visible screen and performing taps, text entry, scrolling, toggles and navigation. Also exposes screenshots, layout, console logs, network requests and hardware keys, and attaches new JS errors and failed requests to the action that caused them.