Preset actions
Preset actions are isolated steps that perform a single action on the current page and succeed immediately when that action is complete. They are the most granular steps available on Momentic, and serve as the building blocks for AI actions.
In general, preset actions can be divided into two categories:
- Interactive: Actions that interact with elements on the page, such as clicking on an element, selecting from a dropdown, or typing some content into an element.
- Browser: Actions that affect the browser itself, such as hitting the back button, switching tabs, or refreshing the page.
Preset action types
The following table lists the available preset action types:
| Name | Type | Description | Options |
|---|---|---|---|
Click | Interactive | Locate an element on the page matching a descriptor, and then click on it. The element must be actionable for the click to succeed: it cannot be hidden, disabled, or animating. | Double click: double click on the element.Right click: Right click on the element. |
Press keys | Interactive | Press a sequence of keys on the page. Multiple keys can be pressed simultaneously by concatenating them with + (e.g. CTRL+A). Key presses may be used to invoke browser shortcuts. | None |
Type | Interactive | Execute a Click command followed by a Press command. The element targeted should be an input (opens in a new tab) element. | Press enter: press the enter key after typing completes (can be used to submit forms easily).Clear existing: clear any existing content in the text input before typing. |
Select option | Interactive | Select an option from a dropdown menu. The element targeted should be a select (opens in a new tab) element. | None |
Wait | Browser | Wait for the specified number of seconds. This step type should be used rarely since other actions already automatically wait for network requests to complete, up to a maximum of 8 seconds.If you need to wait for a particularly long request to complete, or if you expect the DOM to change after a large initial delay, you can use the Wait action to ensure Momentic actions on the correct state of the page. | None |
Scroll up | Browser | Scroll up by a single page height. | None |
Scroll down | Browser | Scroll down by a single page height. | None |
Go back | Browser | Return to the previous page in the browser history. | None |
Go forward | Browser | Advance to the next page in the browser history. A Go back step must be invoked prior to this action. | None |
Navigate | Browser | Go to the specified URL. | None |
Switch tab | Browser | Switch to the tab with the specified URL. This action can be useful for following pop-ups or redirects that open in a new tab or window. The input value is used to substring (opens in a new tab) match the URLs of each available page. The first page that matches is set as active. | None |
Set cookie | Browser | Set a cookie that will persist throughout this browser session. The cookie should follow the HTTP Set-Cookie (opens in a new tab) format. This action can be useful for reusing sessions and bypassing the need to re-authenticate, however users must manage cookie expiry. | None |
Best practices
Using the locator button
The locator button is available on commands that target a specific element
(Click, Type, Select)
The eye-shaped locator button allows users to preview what element would be targeted by an interactive step. When pressed, Momentic's AI web agent uses the user-provided descriptor to find a matching element on the page, and then highlights that element. If no element matches, an error message is displayed on the step card.
This feature can be used to "dry-run" actions and ensure that the correct element will be targeted during a test run. If Momentic highlights the wrong element, you can adjust your element descriptor without having to re-run all steps (view our guidelines for writing accurate descriptors here).
Additionally, for Select actions, the locator button populates the available dropdown values in the test editor. Using the locator button is highly recommended for Select steps to avoid the possibility of misspelling the desired dropdown value.
Writing element descriptors
The first input field on interactive commands is always an element descriptor. This descriptor is passed to Momentic's AI web agent to locate a specific element during test runs. We recommend following the guidelines below to craft descriptors that are easily understandable by our web agent:
Be specific about element type
Describe the desired element with a specific web accessibility role (opens in a new tab), such as "the submit button", "the homepage link", or "password input". If you are unsure what type of element you are targeting, you can use the Chrome Dev Tools (opens in a new tab) to find out.
Include labels
If the desired element has a label, include that in your descriptor, such as "the search button" or "the username textbox". Our AI web agent will generally attempt to match this label to the "name", "id", or "aria-label (opens in a new tab)" attribute of HTML elements on your website. Defining ARIA attributes not only improves the performance of Momentic tests, but also helps your website stay compliant with ADA standards (opens in a new tab).
Include positioning cues
Our AI web agent understands basic relative positioning, and so you can provide cues such as "...below the 'Introduction' header" or "...beside the 'Back' button". Only use positioning cues if they are expected to be stable over time.
Avoid anti-patterns
The following is currently unsupported by our AI agent:
- Targeting an element using its appearance (e.g. color or size).
- Targeting an element that is
hiddenby CSS or blocked by another element. - Targeting static elements, such as raw text (
<h1>) or images (<img>). If a static element appears to be interactive on your website, it is likely because there is a hidden link (<a>) above or around your element and you should target that link instead. You can still write assertions about static elements.