AI Agents
Agentic mobile testing on simulators. Give an AI coding agent the Appetize CLI and let it run, drive and debug your iOS and Android apps.
Agentic mobile development on simulators: give an AI coding agent the Appetize CLI and let it drive your iOS or Android app.
The appetize CLI runs your iOS or Android app on an Appetize device and drives it from your terminal. Sessions are headless (no browser, no embed), so the same commands work on your machine and in CI.
Beta. Command names, flags and output are all still changing between releases.
npm install -g @appetize/cli
appetize skill installNode 22 or later is required. Confirm the install with appetize --version.
Let an agent drive
An agent can take a whole task end to end. Investigate, fix, verify, ship:
"Take this crash log. Validate it, use the app and network logs to find the root cause, fix it, test it and ship it."
"Ship the new sign-up copy. Update it, check it on an iPhone 15 Pro and a Pixel 7, and put the screenshots on the PR."
"Onboarding regressed. Walk it, find where it breaks, fix it, and send me a recording of it working."
Every step is a terminal command, so it loops on its own: change the code, re-run on a simulator, check the recording, repeat. A save can also update the running app, with no rebuild. See AI Development Loop.
Or drive it yourself
Every command the agent runs, you can run by hand. That is what makes an agent's session debuggable: when it says the tap missed, you can start the same session and look.
Next steps
Getting started: install, token, and your first agent task
Sessions: lifecycle, logs and network traffic
Automation: inspect the screen and act on it
Screenshots and recordings: capture PNGs and MP4s
AI Development Loop: save, look, and edit again while the app stays open
Last updated