Scrape data from several web pages with Automatio Chrome AI

Ask the extension to visit a list of addresses or find links on a page to extract specific information. The AI agent opens its own tabs in your browser to read, scroll, and collect the data you need.

What to ask

Type it in the Automatio side panel in Chrome, with the page open. Send it as written, or change the details to fit your case.

Go to [URL] and find the links for the five most recent blog posts. Open each link in My browser mode and extract the headline, the author name, and the main summary of the content. Display the results in a table.

How it works

  1. 1

    Open the side panel

    Click the Automatio icon in your Chrome toolbar to open the side panel and sign in to your account.

  2. 2

    Set the agent mode

    In the chat composer, open the agent menu and select My browser so the agent can work within your signed-in Chrome session.

  3. 3

    Send your request

    Type your scraping instructions into the chat, including the starting URL and the specific data points you want to extract from each page.

  4. 4

    Monitor the agent

    The agent will open new tabs to perform the work; ensure you keep a page of automatio.ai open while it runs and do not dismiss the debugging bar.

What you get

The agent returns the extracted data directly in the side panel, often formatted as a table or a structured list. You should review the text for accuracy against the source pages before copying it or saving the result. If the agent encountered errors on specific pages, it will report those alongside the successful extractions.

Since the agent operates in your browser, it can access pages you are already signed into, provided it opens those tabs itself.

Good to know

  • The agent only works in tabs it opens itself and cannot scrape your existing open tabs.
  • Tools built to scrape multiple pages are limited to opening at most 50 tabs per run.
  • The extension cannot scrape data from chrome:// pages or the Chrome Web Store.
  • Using the agent or prompt tools consumes credits from your Automatio account.

Variations to try

Scrape from a list

I will provide a list of 10 URLs. Open each one in My browser mode and tell me if the phrase [Keyword] appears in the first two paragraphs of each page.

Extract and download

Go to [URL], find the links for all PDF whitepapers, and download each one to my computer. Limit this to the first 5 files found.

Scroll and scrape

Open [URL], scroll down three times to load more items, and then extract the names and prices of every item visible on the page.

Questions

Can it scrape pages that require a login?

Yes, if you use My browser mode, the agent operates in your signed-in Chrome session. As long as the agent opens the tab itself, it can access content you have permission to view.

Does it work in Firefox or Safari?

It does not. The extension requires specific APIs only available in Chromium-based browsers like Chrome, Edge, and Brave.

Is there a limit to how many pages I can scrape?

The AI agent can navigate multiple pages as requested, but script-based tools are limited to opening 50 tabs per run. Each action taken by the agent uses your account credits.

Can I scrape data from a PDF?

The @page command cannot read PDFs in the Chrome viewer. You must use @screenshot to capture the visible part or @url to let the agent attempt to read the file directly.

More guides