web-search

To fetch real-time information from the web, you can search current events and locate specific API documentation by querying the web and retrieving extracted markdown content from authoritative pages. This approach retrieves verifiable answers not available in training data.

To extract full text content from a specific URL into markdown, you fetch the target page after an initial search. This converts the URL's content into readable markdown so you can read details beyond the initial search snippets.

Yes, you can run multiple web searches concurrently. You can execute up to 10 concurrent queries to compare versions, vendors, or policies, returning summarized answers and ranked result pages for broader coverage.

The best way to verify facts and find current API documentation online is running a targeted web search to locate result pages, then explicitly fetching the most relevant official documentation URL to verify the exact conditions and extract its full markdown text.

A limitation of fetching web content is that it requires a two-step process: an initial search step returning result pages, followed by an explicit URL fetch to extract the full markdown content, meaning you cannot retrieve full page text directly from a single search query.