Mobile Agent Credits
The Mobile Agent uses credits. Your plan gives you a set number of credits each month, and each agent run uses some of them.
An agent run is one prompt the Mobile Agent runs. Each agent run uses 2 credits.
Action | Credits used |
Run a prompt (type a task and click Send) | 2 |
Run a Saved Prompt | 2 |
Prompt in a schedule | 2 per prompt, each time the schedule runs |
Failed or stopped run | 2 |
Convert a run to a No-Code test case | 0 |
The cost doesn't depend on how long the prompt is or how many steps the agent takes. A one-line task and a long multi-step flow each use 2 credits.
A failed run, or one you stop before it finishes, still uses 2 credits.
Scheduled prompts: each prompt in a schedule uses 2 credits every time the schedule runs. For example, a schedule with 5 prompts uses 10 credits per run.
Check your Credit Balance
Your balance is shown at the top right of the Mobile Agent page, above the prompt box.
For example, 28 / 1500 Credits · DEVELOPER means:
- 28: credits used this month
- 1500: your monthly credit allocation
- DEVELOPER: your plan
Monthly reset
Your credits reset every month to your plan's allocation.
Running out of credits
When you run out of credits, you can't run any more prompts with the Mobile Agent until your credits reset.
To get more credits, do either of the following:
- Email customersupport@sofy.ai
- Click Submit Ticket at the top of the page
Tips to get the most from your credits
- Be specific. A clear prompt is more likely to succeed on the first try, so you won't need to spend 2 more credits re-running it. Name the screens, buttons and expected result, for example: "Log in with the saved test account, open Settings, and verify that the Notifications toggle is on."
- Check the device, OS and build in the selectors below the prompt box before you click Send. A run on the wrong setup still uses 2 credits.
- Use Saved Prompts for tasks you repeat, so the wording is the same every time.
- Watch your schedules. Prompts in schedules use credits on every run, so check how many prompts each schedule has and how often it runs.
- Convert stable flows to No-Code test cases. Converting a run uses no credits.
