How to Configure a Browser Agent to extract data

Bardeen’s Browser Agent Builder allows you to configure an agent that can extract information from a webpage. It’s fast and intuitive: you select elements on an example page, and Bardeen will repeat the logic on similar pages.

You can configure agents to extract:

• Specific fields from a single page
• Repeating information from lists
• Structured tables
• Or even deep crawl by opening sub-pages (see Deep Crawling with Agents)
• You can also build agents that extract multiple lists from one page. Learn more in How to configure an agent for extracting multiple lists? 

📄 Single Page Extraction

A single page agent extracts static elements like text, links, or images from into a single row of data.

Start Building

1. In the Scrape data on active tab (or in the background) action, click + Create Scraper.
2. Select or paste a link to an example page.

⚠️ Important:
This URL becomes the template. You can't change it later without rebuilding, so choose a page that contains all data fields you want to extract.

Select Elements

1. Click on highlighted elements and choose Get text, Get link, etc.
2. Name the field (e.g., “Company Name”, “Industry”).
3. Repeat for each item.

🖱️ Advanced Actions

Depending on how the page behaves, you may need to interact with it to extract information.

• Click – Expand or trigger elements.
• Input – Enter values or fill forms.
• Get Image – Extract image URLs.
• Hover – Reveal hover-only content.
• Get HTML – Capture raw HTML of an element.

To remove or rename a field, use the menu on the right side. You can also add extra logic using the + button (see Special Elements).

➕ Special Operation

To refine your agent logic, you can use additional actions by clicking the ➕ button in the Builder, those are the actions that doesn't require specific element on a page.

• Delay – Wait before the next action. Useful after a Click.
• Meta image / title / description – Extract page-level metadata.
• Page URL – Add the current page’s link.
• Clipboard – Access copied content.
• Time scraped – Timestamp of the run.

📋 List Extraction

List operations allow extracting data from repeating sections like product listings or search results from a website.

Create the List Logic

1. Open the page with the list.
2. Click on a single item (e.g., one result).
3. Choose List to enter list mode.
4. Click on another similar item to define the structure.

Note:
You only need to define the structure with two examples — Bardeen will repeat it for the rest.

Configure Pagination

If there are multiple pages:

• Click pagination – Pick a “Next” or “Show more” button.
• Infinite scroll – Auto-loads on scroll.

After current step you should see all the relevant items highlighted:

Now you can start picking items inside of the list, similar to how you would do in Single Page Extraction:

But now this action will be repeated for every single list item and extract every single item from the defined list:

🧾 Table Extraction Agent

If a webpage contains a table, Bardeen can auto-detect and extract it with structure.

1. Click any cell inside a table.
2. Choose Table from the menu.
3. To extract additional elements like links, click Activate and then select more fields.

🧾 Editing Browser Agent Template with a List

When editing agent model and adding new fields to existing list item like in the image below:

You might notice that the item in certain situation is being added as a separate field:

This is because the lists are disabled by default and every item added is treated as single page item so you can combine both single and list operations, but if intend to add item to the list then we need to activate it first by clicking on activate icon:

⚙️ Agent Settings

Click the gear icon in the Builder to configure options.

Where should the agent run?

• Everywhere – Use globally.
• On current domain – Recommended default.
• Custom – Use custom logic or regex.

Other settings:

• ✅ Use minimized window – Turn off for sites like LinkedIn that block minimized windows.
• ✅ Scrape in the cloud – Works only on public pages. Not supported for gated sites.

👇 What’s Next?

• 🔍 How to Configure a Deep Crawling Agent – Navigate into subpages and extract more layers of data.
• 🧩 How to configure an agent for extracting multiple lists?