Iphone Mirroir Mcp

Bridges macOS iPhone Mirroring to MCP clients so agents can see, tap, and type on a real iPhone. Uses Apple Vision OCR by default to pull text and coordinates from screenshots, with optional YOLO CoreML for icon detection and embedded AI vision for semantic UI understanding. Exposes 33 tools covering basics like tap, swipe, type_text plus navigation helpers like scroll_to, press_back, reset_app, and set_network. Works with any MCP client over stdio. Requires macOS 15+ and grants for Screen Recording and Accessibility. Ships with a skill system that lets you describe your app's structure in APP.md files to guide exploration. Install via npx, Homebrew, or the curl script at mirroir.dev.