Most AI coding agents can write mobile code. The difficult part begins when the app is running. An agent may know what the code is supposed to do, but it often cannot directly answer: - What request did the app actually send? - What response did it receive? - Did the request fail before reaching the server? - What data was persisted in SQLite? - Did the database contain the row the app expected? Without that runtime context, mobile debugging becomes guesswork. I built Lynx to close this gap. Lynx is a vendor-neutral, read-only CLI that gives developers and AI agents direct visibility into a debuggable mobile app: 🌐 Capture HTTP/1.1, HTTPS, HTTP/2, and WebSocket traffic 🗄️ Inspect SQLite schemas and run read-only SQL queries 📱 Attach to Android devices, Android emulators, and iOS Simulators 🤖 Return structured JSON that an AI agent can inspect and reason about ⚡ Run as a native Kotlin Multiplatform executable The goal is simple: give AI agents the same runtime facts a developer needs when debugging a mobile application. A typical workflow looks like: attach → start capture → inspect network traffic attach → inspect database → run a read-only query Lynx is open source and still evolving. I am especially interested in feedback from Android, iOS, Kotlin Multiplatform, and AI tooling developers. 🔗 https://github.com/RotBolt/lynx

lynx

Give AI agents eyes into mobile network and SQLite data

Give AI agents eyes into mobile network and SQLite data