Skip to main content

Browser

Browser placeholder

Browser allows Subagents to navigate and read page content when answers require targeted web extraction.

When to use this tool​

  • Website content extraction
  • Public page verification
  • Dynamic lookup from known URLs

Configuration​

  • Specify trusted domains/sources
  • Configure auth if protected pages are required
  • Define extraction boundaries

Browser config placeholder

Example usage​

User prompt:

Open page after searching Stochastic.ai

Expected behavior:

  • Accesses specified page(s)
  • Extracts requested details
  • Returns concise structured findings

Browser example placeholder

Validation checklist​

  • Confirm tool is attached to correct Subagent
  • Test both success and failure path in Preview
  • Verify deployed behavior in Interact for real user phrasing

Common issues​

  • Restricted pages fail without auth config
  • Broad extraction requests return noisy output

Best practices​

  • Constrain to known domains
  • Ask for specific fields instead of open-ended scraping