Built an Autonom — a mobile test & debug harness built for AI coding agents.
Gives Grok, Claude and Codex real control of Android emulators and iOS Simulators:
→ identical accessibility tree schema on both platforms
→ semantic find / tap / type
→ full action journal + screenshots + logs
→ consent-gated network capture & mocking
Try it and share your feedback!
https://t.co/2TKigsrWR5
Hard-to-reach screens without fighting the UI
```
autonom session start --serial emulator-5554 --app-id https://t.co/TV8UhZaBEz
autonom open "myapp://order/42"
autonom permissions grant photos
autonom location set 55.751244 37.618423
autonom media add ./fixtures/photo.jpg
autonom ui tree
autonom screenshot
```
Agent sets deep link + permissions + location + media in one flow, then immediately verifies the resulting screen with the accessibility tree and a screenshot.
No more manually navigating through system dialogs or guessing coordinates.
Built an Autonom — a mobile test & debug harness built for AI coding agents.
Gives Grok, Claude and Codex real control of Android emulators and iOS Simulators:
→ identical accessibility tree schema on both platforms
→ semantic find / tap / type
→ full action journal + screenshots + logs
→ consent-gated network capture & mocking
Try it and share your feedback!
https://t.co/2TKigsrWR5
Error paths without touching the backend:
```
autonom network mock add --url "*/login" --status 500
```
Agent taps Login → verifies the error state in the accessibility tree → exports HAR (credentials already redacted).