Browse websites
Give your agent the steps you’d take in a browser:- Open a website.
- Open another browser window when the task needs more than one page.
- Go back or forward through browser history.
- Wait for a page to load before taking the next step.
- Check basic details about the current browser window.
Interact with pages
Tell your agent what to click, type, or look for:- Click a button, link, or other page element by describing it, like “the Search button.”
- Type into a field or fill out an entire form.
- Hover over an element to reveal a menu, tooltip, or other content.
- Scroll through the page.
- Wait for a change you describe, like “the results table appears.”
Get information from pages
Turn what your agent reads into information it can use:- Ask a question about the current page and use the answer in the next steps.
- Use AI to make sense of information across the pages in a browser session.
- Collect names, prices, links, or table rows in a structured format you can use in later steps.
Work with integrations
Connect your browser tasks to the apps you use:- Ask a connected app to take an action or get information, using plain English.
- Pass information between a website and a connected app as part of the same agent’s run.
Available apps and actions depend on the services you’ve connected. Visit integrations for the full list of services you can connect with Airtop.
Use downloaded files
- Wait for a file download to complete.
- Retrieve the file so your agent can use it in later steps.
Put it together
Combine these capabilities to build an agent that:- Visits a website, collects the information you need, and sends it to a connected app.
- Fills out an online form with information you provide.
- Watches a page for a change and takes the next step when it happens.
- Starts a download and processes the file once it’s ready.