minhngoc25a Tavily MCP map
Use mcp__tavily-mcp__tavily_map to discover URLs on a site without extracting page content.
When to use
- You know the site but not the exact page.
- You need a list of URLs before deciding what to extract or crawl.
- The user asks for site structure or pages under a domain.
Use tavily-mcp-extract after map when only a few discovered pages are needed. Use tavily-mcp-crawl when many pages need content extraction.
Parameters
| Parameter | Use |
|---|---|
url |
Root URL to map. |
instructions |
Natural language guidance for which URLs to return. |
max_depth |
How many levels from the root to explore. Start low. |
max_breadth |
Maximum links to follow per level. |
limit |
Total URL cap. Set this explicitly for larger sites. |
select_domains |
Regex patterns for allowed domains or subdomains. |
select_paths |
Regex patterns for included paths. |
allow_external |
Whether to include external links. Disable when only the target site matters. |
Patterns
Find a docs page
Map the docs root with instructions describing the page or topic, then extract the best URL.
Inspect site structure
Map the root with conservative max_depth and limit, then summarize URL groups.
Focus on a section
Use select_paths for sections such as /docs/.*, /api/.*, or /guides/.*.
Tips
- Map returns URLs only; it does not read page content.
- Map plus extract is often better than crawl when only a few pages are needed.
- Start with
max_depth: 1and a modestlimit, then expand if needed. - Use
instructionswhen URL patterns are not enough to express the target.