Releases·Tool Release
Dev adds iPhone Duo support to MobAI, uses Claude Code to test iOS app across fold states
A developer added iPhone Duo support to MobAI, a desktop tool that lets coding agents control mobile devices, and used it with Claude Code to test an iOS project for the foldable. Claude drove the simulator across fold states and produced 17 findings. It shows coding agents doing hands-on UI testing on unreleased hardware.
4 engagement1 sourceSun, Sep 20, 2026, 08:46 PM
AI Overview
- MobAI lets coding agents tap, swipe, and set hinge angles on the iPhone Duo simulator; it is free for one device, no account or telemetry
- A 12-minute Claude Code run exploring the app across fold states returned 17 findings on layout, accessibility, and behavior
Entities
Related
Products·building
Claude cloud agent controlling iPhone without Mac
A developer is building a device-control layer that lets a Claude cloud agent remotely control a physical iPhone. The agent runs in Claude's sandbox and interacts with the phone in real time, no Mac or cable needed. It solves the problem of needing a Mac to run iOS automation or testing agents.
A developer is building a device-control layer that lets a Claude cloud agent remotely control a physical iPhone. The agent runs in Claude's sandbox and interacts with the phone in real time, no Mac or cable needed. It solves the problem of needing a Mac to run iOS automation or testing agents.
99 engagement1 sourceredditThu, Jul 30, 2026, 06:02 PM
Social·Community
If your coding agent is building & testing a mobile app, how are you thinking about emulators/ simulators vs real devices strategy?
A developer using Cursor across teams to run a full agent loop (code, run app, write tests, inspect failures, fix, iterate) asks how peers give coding agents access to mobile environments. Web agents can work against a browser directly, but mobile requires emulators, simulators, or physical devices. The thread seeks real-world setups and whether agents have shifted simulator-vs-device reliance.
A developer using Cursor across teams to run a full agent loop (code, run app, write tests, inspect failures, fix, iterate) asks how peers give coding agents access to mobile environments. Web agents can work against a browser directly, but mobile requires emulators, simulators, or physical devices. The thread seeks real-world setups and whether agents have shifted simulator-vs-device reliance.
2 engagement1 sourceredditTue, Oct 6, 2026, 07:55 PM
Releases·Tool Release
Claude MacOS Desktop adds native iOS Simulator control for developers
Anthropic's Claude MacOS Desktop app now supports direct control of the iOS Simulator, enabling automated app testing and development workflows. The feature is aimed at iOS developers seeking to integrate AI-assisted testing into their toolchain.
Anthropic's Claude MacOS Desktop app now supports direct control of the iOS Simulator, enabling automated app testing and development workflows. The feature is aimed at iOS developers seeking to integrate AI-assisted testing into their toolchain.
134 engagement2 sourcesreddit, rssWed, Jul 22, 2026, 08:53 AM
Products·launched
Metro Remote: test AI-generated React Native code on real iPhone
Metro Remote lets developers test AI-generated React Native code on a real iPhone over WiFi or 5G. After an AI agent (Claude or Codex) writes a code fix, Metro Remote applies it via Fast Refresh, then the agent reruns the broken flow on the physical device to verify the fix works. It solves the problem that AI coding tools stop at writing code, but mobile apps must be proven on real hardware.
Metro Remote lets developers test AI-generated React Native code on a real iPhone over WiFi or 5G. After an AI agent (Claude or Codex) writes a code fix, Metro Remote applies it via Fast Refresh, then the agent reruns the broken flow on the physical device to verify the fix works. It solves the problem that AI coding tools stop at writing code, but mobile apps must be proven on real hardware.
3 engagement2 sourcesredditMon, Jul 13, 2026, 10:00 PM
Releases·Tool Release
I built an app to use Claude Code from my phone and test the result there too (simulator, web, API)
A developer released CosmoRemote, a native iOS/Android app with a Mac bridge daemon that lets users drive Claude Code from a phone and test the resulting builds remotely. Unlike Anthropic's own remote Claude Code offering, it streams the simulator, web preview, and local API to the phone so users can verify changes before sending the next prompt. It targets the workflow gap where Claude reports 'done' but the user can't build or run the change away from their desk.
A developer released CosmoRemote, a native iOS/Android app with a Mac bridge daemon that lets users drive Claude Code from a phone and test the resulting builds remotely. Unlike Anthropic's own remote Claude Code offering, it streams the simulator, web preview, and local API to the phone so users can verify changes before sending the next prompt. It targets the workflow gap where Claude reports 'done' but the user can't build or run the change away from their desk.
5 engagement1 sourceredditSun, Sep 20, 2026, 06:05 PM