AI Magic

AI Magic lets you describe a task in plain English and have the agent figure out how to complete it. Instead of specifying every click and keystroke, you give a high-level instruction and the agent handles the details.

How It Works

Add an AI Magic step to your test case and provide a natural language instruction. The agent interprets the instruction and performs the necessary interactions.

Configuration:

• instruction (required): Natural language instruction for the agent to execute

Examples:

• "Sort products from cheap to expensive"
• "Select next Friday on the date picker"
• "Fill out the shipping form with a random address"

The more specific your instruction, the more precise the execution. "Add the first blue t-shirt in size medium to cart" works better than "Add to cart".

Calendar Interactions

Date pickers and calendars are usually a testing nightmare. Tell the agent to "select next Friday" or "book a meeting for December 15th at 2 PM" — it handles all the clicking, scrolling, and month navigation automatically. Whether it's a custom React component, a jQuery plugin, or native HTML5 date inputs, the agent figures it out.

Extract and Reuse Data

The agent can grab information from the page and save it for later. Extract prices, order numbers, user names — anything visible on screen — and store them as session variables.

• Compare prices across different pages
• Verify order confirmations match cart totals
• Check that usernames appear correctly after login

Say "save the order total as checkout_amount" and use it later in assertions or other steps.

When to Use AI Magic

AI Magic is best for multi-step interactions that would require several individual steps — date pickers, complex form fills, multi-click workflows. For simple single actions (clicking a button, typing in a field), use the dedicated step types instead.

See the full Test Steps & Actions reference for all available step types.