Commands

Closed Beta: The Command-Based Step Execution is currently under development and only available to a selected group of companies. Currently available only for Android, the feature is expected to launch for iOS in Q3 2025. Feedback is highly welcome.

WebView automation through commands currently has limited support. We recommend using natural language (AI) steps for WebViews instead.

The Command-Based Step Execution feature allows for efficient test execution by running commands directly, without initially relying on AI. This approach significantly improves run times and reduces costs. AI is only used as a backup when unexpected elements appear, such as marketing popups, notifications, UI changes, text changes, or missing element IDs.

To get started with Commands, check out this video showing how our new Recorder feature works — it makes creating command-based tests quick and easy

We currently support these gestures:

• Tap Command

• Type Command

• Scroll Command

• Scroll Down Until Visible

• Scroll Up Until Visible

• Assert Visibility of an Element ID

• Assert Visibility of Specific Text

• Detecting Audio Playback

• Executing a cURL Request

• Execute Network Call by Selector

• Navigate to a Specific Screen

• Launch App by ID

• Long Press on Element by Text

• Long Press on Element by ID

• Match Biometry

• Open Deep Link URL

• Press Android Back Button

• Remove Text

• Restarting the App

• Set Location

• Shake Device

• Verify Email

• Swipe Command

• Slide Command

• Wait Command

Distinguishing steps executed using the command-based steps and the LLM/ AI

You can identify command-based steps by opening the step in the execution log and checking the ‘Reasoning from GPT Driver’ If no values are provided, the step was executed without AI assistance and is considered command-based.

A command-based execution

If ‘Reasoning from GPT Driver’ is provided, the step was executed by the AI.

A LLM guided execution