Browser automation

Some data lives behind interaction — a button click, a form, a multi-step flow. The String Web Access API gives you two ways to handle it.

Lightweight: browser actions

For short, scripted sequences (and optional screenshots), pass an actions array to /fetch. The API runs them in a managed browser and returns the final HTML and any screenshot.

{
  "url": "https://example.com",
  "actions": [
    { "type": "click", "selector": "#accept-cookies" },
    { "type": "write", "text": "hello" },
    { "type": "press", "key": "Enter" },
    { "type": "screenshot" }
  ]
}

When the data you want arrives from an API call rather than in the HTML, end the sequence with a waitForResponse action and asResult: true. The API response comes back as data, so there is no page to parse. See XHR capture.

When the form or widget you need is inside an iframe, give the action a frame to run it there, and add includeIframes: true to get the iframe's content back with the page. See Iframes. When a click sets off requests you can't name one by one, follow it with a waitForNetworkIdle action. See Waiting for page load.

Full control: CDP browser

For open-ended automation (logins, long flows, anything dynamic), connect Puppeteer or Playwright to the CDP browser and drive the page directly.

Choosing between them

Both apply access control to every navigation.