Enable optional content extraction, scrapping in WebSearchTool

[priyanshuTelus]

Hi Team,

Thanks for the great work on the Agents SDK — it's been fantastic to work with.

Currently, WebSearchTool returns only metadata: title, snippet, and URL. It does not extract or fetch content from the linked pages. This limits its usefulness for real-time, information-rich tasks like:

• Wildfire or weather updates
• Traffic incidents or road closures
• Government alerts or advisories

In these cases, the links are helpful, but agents still need a secondary scraper to retrieve usable data.

• Proposed Solution
  Add an optional parameter to fetch and extract main page content:
  WebSearchTool(fetch_page_content=True, content_limit=5000)

Support for basic article extraction (e.g. Readability) and optional JS rendering (e.g. via Playwright) would significantly enhance the tool’s utility.

• Enables content-aware agents out-of-the-box.
• Reduces need for external scraping tools.
• Unlocks more real-time use cases.

Thanks for considering this!

[seratch]

Thanks for sharing this idea. Unfortunately, our platform does not have immediate plans to enhance this way, so please go ahead with your own function tool, which does these.