Getting started
What can it do?
What can it do?
What do I need?
What do I need?
Can Claude control my iPhone?
Can Claude control my iPhone?
Which agents work with it?
Which agents work with it?
mobster mcp, mobster run --json or the local HTTP API. Add Mobster to your agent has the setup for Codex, Cursor, VS Code and nine more clients. Your agent can drive the phone itself, or hand a whole task to Mobster’s agent with phone_task. Claude on the web or your phone, ChatGPT, Poke and other assistants in the cloud connect through Mobster Remote.How is it different from TapKit, Maestro or Appium?
How is it different from TapKit, Maestro or Appium?
Does it work on Android?
Does it work on Android?
Which iOS versions?
Which iOS versions?
Your iPhone
Why does it need Developer Mode and Xcode?
Why does it need Developer Mode and Xcode?
Does it work over Wi-Fi?
Does it work over Wi-Fi?
Can I control a phone that isn't connected to this Mac?
Can I control a phone that isn't connected to this Mac?
Can it run on a schedule?
Can it run on a schedule?
Models and costs
What does a task cost?
What does a task cost?
Does it remember things about me?
Does it remember things about me?
Where do my screens go?
Where do my screens go?
Can I attach files or talk to it?
Can I attach files or talk to it?
Approvals and safety
Is it safe?
Is it safe?
Can it unlock my iPhone?
Can it unlock my iPhone?
Plans and refunds
How much does Mobster for Mac cost?
How much does Mobster for Mac cost?
Is there a free trial?
Is there a free trial?
Do I need an account?
Do I need an account?
How many Macs does a plan cover?
How many Macs does a plan cover?
Does Mobster for Mac contact your servers?
Does Mobster for Mac contact your servers?
Can I cancel a monthly or yearly plan or get a refund?
Can I cancel a monthly or yearly plan or get a refund?
For developers
I build iOS apps. What's in it for me?
I build iOS apps. What's in it for me?
mobster test runs your project’s checks on simulators Mobster manages and on your own iPhone, with retries, flaky checks named, and JUnit and HTML reports for CI. Each verdict comes from assertions on the app’s accessibility tree, never from a model, and your coding agent runs the same checks with run_tests. There’s no per-device fee. Run every checkDo the checks heal themselves?
Do the checks heal themselves?
mobster test replays its recorded steps with no model. When a button moved or was renamed, Smart redoes that one step and the change waits in <check>.heal.yaml for you to accept or reject. A heal never changes an expectation, so a real bug still fails. --no-heal fails the check instead. See Replay and heals.Will it touch my own simulators?
Will it touch my own simulators?
Mobster · …, and deletes only those.