User Actions | Applitools Documentation
The Autonomous Natural Language Processing (NLP) Engine
The Autonomous Natural Language Processing (NLP) engine allows you to effortlessly create a custom flow test using simple, everyday language. This powerful technology lets you define steps using common English phrases like "Browse to http://example.com/" or "Click the OK button" without any coding required. You can include variables or parameters in a test to allow you to run the test in different scenarios.
As you enter instructions for a step, Autonomous provides helpful suggestions to guide you through the process.
Auto-Correction
The system automatically corrects syntax errors and divides complex steps into multiple simple steps, making them easier to read and understand. The Auto-Correct feature makes the following corrections to commands:
- Grammar and Spelling – For example, "Clack OK" will automatically be corrected to "Click OK".
- Break Down Complex Steps – Steps with multiple actions will be broken down into shorter, more manageable steps. For example, "Enter 'John' in 'User name' and click OK" will be divided into:
Enter 'John' in 'User name'Click 'OK'
- Simplify and Standardize Commands – Steps will be modified to ensure consistency and clarity across all steps. For example, "In the user name field, enter 'John Smith'" will be changed to "Type 'John Smith' in the 'User Name'".
If steps were corrected, the
For more on what PDF testing can do, see PDF Testing.
Entering Instructions
When entering an action, identify a screen element, such as a button, text box, or image, and then specify what to do with that element, such as click, hover, or enter text.
Elements
Elements on the page can be identified in various ways, including:
- Text label, for example, "Get Started" button.
- Position in a list, for example, "The third item in the Language list".
- CSS Selector or X-Path selector ID, for example, "gh-minicart-hover".
- Position in relation to another item, for example,
Checkbox to the right of "I agree". - Location on the page, for example
button at the top of the page.
Screen elements with more than one word should be in quotation marks ("...").
Common Commands
Here are examples of common commands that the NLP engine supports. This is not a complete list, as there is no limit to the types of commands you can create.
| Command | Definition | Examples |
|---|---|---|
| - Click | Click on a screen item identified by text, location, or a CSS Selector or X-Path selector ID | - Click "Advanced"- Click "gh-minicart-hover"- Click the "OK" button- Click the first button- Click the "Shopping Cart" icon- Click the button below "Are you sure?"- Click the second radio in the "Preferred Contact Method" radio group- Click on the "SMS" radio in the "Preferred Contact Method" radio group |
| - Type - Enter |
Types text in a field | - Type "Jane Smith" in the "Your Name" field- Enter "janesmith@example.com" in the "Email" field |
| - Select - Choose |
Select an item from a drop-down list | - Select the third option in the "Country" field- Choose "United States" in the "Country" field- Choose "English" from the "Language" dropdown menu |
| - Go to - Navigate to - Browse to |
Opens a URL. The system automatically enters this as the first step in the test. | - Go to "https://applitools.com/platform/eyes/"- Navigate to "https://applitools.com/solutions/mobile-testing/"- Browse to "https://applitools.com/about/" |
| - Go to [PDF page] - Navigate to [PDF page] - Jump to [PDF page] |
Navigates to a specific page in a PDF document opened as the test URL. Supported for PDF testing only. Ordinal words (first, second, last) and numbers (3rd, 5th) are both accepted. For web page navigation, use Go to / Navigate to / Browse to. |
- Go to the first PDF page- Go to the third document page- Go to the 3rd PDF page- Navigate to the second page- Go to the last PDF page- Jump to the 5th page- Switch to the second document page |
| - Scroll | Scrolls the screen up, down, left, or right. | - Scroll to the top of page- Scroll left- Scroll down 30 pixels- Scroll to the "OK" button- Scroll down 2 times |
| - Check - Uncheck |
Selects or clears a checkbox. | - Check "I agree"- Check the second box- Uncheck the checkbox to the right of "I would like to receive updates"- Uncheck "Keep me signed in" |
| - Hover | Positions the pointer over a screen element; this is useful to capture a tooltip or to select an item from a popup menu. | - Hover over "Company"- Hover over "gh-minicart-hover"- Hover over the first link |
| - Clear | Removes text from a field. | - Clear the "First name" field |
| - Press - Hit - Enter |
Presses a keyboard key. | - Press ENTER- Press ESCAPE- Press backspace- Press delete- Hit TAB- Press the left key |
| - Refresh - Reload - Soft refresh - Hard refresh |
Refreshes the page. A hard refresh clears cached data when the page is refreshed. This is the equivalent of pressing CTRL+F5 | - Refresh the page- Reload the page- Soft refresh the page- Hard refresh the page |
| - Wait - Pause - Delay |
Waits before performing an action. This is useful if it may take time for a page to load or for a page with lazy-load functionality. | - Wait 3 seconds- Pause 5s- Delay half a minute |
| - Wait | Explicitly waits for a specific element or all elements to load. The system times out only after the defined period. | - Wait 30 seconds for "Search" element to appear- Wait up to 2 seconds for "MyElement" to appear- Wait for elements up to 1 minute |
| - Implicit Wait | In all subsequent test steps, implicitly waits for the maximum defined time for an element to become visible before attempting to interact with it or assert on it. | - Set implicit wait time to 20 seconds |
| - Popup Dialog | Performs an action on a popup window that opens over the page | - Click the first link in the dialog- Click "OK" in the popup |
| - Table Interaction | Tables and grid structures can be interacted with using natural language. | - Click the button at the 2nd row of the "Action" column in the "Orders" table- Type "Approved" in the "Status" column of the "Order 1024" row in the "Orders" table- Verify the "Salary" column of the row "Cedric Kelly" of the table equals $433,060 |
PDF Testing
Custom flow tests can validate PDF documents alongside web pages. Use this to verify short, business-critical documents such as invoices, contracts, statements, reports, and confirmation letters.
How It Works
With PDF testing, you navigate to a PDF the way your users do and test it just like a web page — with Visual AI, text assertions, and interactions. When you open a PDF, Autonomous displays it in the interactive browser. The document renders consistently regardless of the browser or device selected for the test, so results aren't affected by viewer-specific differences.
There are two ways to open a PDF in a test:
- As the Start URL – enter a direct PDF URL (for example,
https://your-app.com/documents/invoice.pdf) when you create the test. - By navigating to a PDF mid-flow – a
Go tostep or aClickon a link that resolves to a PDF is routed through the PDF viewer.
What's Supported
- Text assertions on specific pages, for example
Verify that the first page contains "Total Amount". - Page navigation by ordinal or number, for example
Go to the second PDF pageorJump to the 5th page. - Visual checkpoints on the current page using
Visually check this screen. - Form interaction — for PDFs with fillable form fields, you can type into fields and click controls.
- Zoom, fit-to-width, fit-to-page, and Ctrl/Cmd+F search in the interactive browser during authoring.
Masking Secret Test Data
If a step includes sensitive information, such as a password, PIN number, credit card number, or a URL with basic authentication, when you save the test, the sensitive data is stored encrypted on the server and masked in the interface.
Upload File
Autonomous supports file upload steps in custom flow tests. You can select files you have already uploaded during test authoring.
NLP Command Examples
You can also type file upload steps directly into the step text box using natural language.
upload file "contract_signed.pdf"
attach file "employee_data.csv"
select file "project_plan.docx"
Native Validation Support
Autonomous preserves any constraints declared on the file input element. For example, if the input specifies accept="image/png,application/pdf", only files of those types are accepted.
Documentation Search
Start typing to search across all Applitools documentation.