tavily-map

URL discovery without content extraction is done by mapping a site to return links only, speeding up research. This approach applies the map command to list pages using depth, breadth, and semantic filters before crawling specific targets.

Finding specific pages on a large domain is best handled by mapping the site and applying natural-language instructions to filter results. You can restrict results by path or domain patterns to narrow down the exact pages needed before extracting content.

You can limit the URLs returned during site mapping by setting result caps and restricting by path or domain patterns. These targeted mapping controls ensure you only receive the relevant page list without exceeding your desired scope.

Filtering discovered URLs by path patterns during domain crawling is supported through regex-based filters and domain selection options. You can apply these restrictions directly in the map command alongside natural-language instructions for semantic filtering.

The Tavily map command supports JSON output for URL discovery, allowing integration into automated workflows. It returns the discovered page links structured as JSON data without extracting the actual page content.

Site mapping should be used instead of full content extraction when you need the right page list from a large domain before crawling. It creates a map-to-extract workflow where you first discover URLs, then extract only the specific pages needed.