What problem does it solve?
Manual iOS app testing on physical devices is slow, repetitive, and difficult to scale, requiring developers to manually tap through screens, read accessibility trees, and verify UI state across hundreds of interactions.
Core Features & Use Cases
- Vision-Driven Device QA: Connects to a real iPhone via USB, captures screenshots, analyzes UI state with vision models, and autonomously performs tap, swipe, and type operations through a CoreDevice IPv6 tunnel.
- SwiftUI Source-Aware Testing: Reads Swift source code and accessibility trees to understand every screen, enabling precise element targeting without brittle coordinate-based scripts.
- Remote Tailnet Execution: Exposes the device over Tailscale with capability-tiered allowlists, letting remote AI agents run QA workflows from anywhere without physical access to the hardware.
Quick Start
Use the ios-qa skill to connect to your iPhone, read the current SwiftUI screen's accessibility tree, and automatically tap the submit button after verifying the form is valid.