Summary
A site's shape in one call.
How many pages a site has and how they are arranged, answered from its published sitemap — the question that otherwise means a crawler and an afternoon.
What it is
A Python server that fetches and parses a website's sitemap and returns it as a tree. Structural questions about any site — page counts, section hierarchy, pages that exist but are not linked from the navigation — are answered from what the site already publishes.
What you get
- get_sitemap_tree fetches a site's sitemap and returns the URL hierarchy
- Nothing is crawled: the answer comes from the sitemap the site publishes for search engines
- Ships prompt templates for Claude Desktop — analyze a website, check sitemap health, extract URLs, find missing content, visualize the structure — each taking just a URL
Requirements
Uv, and `uvx sitemap-mcp-server` in the client config with TRANSPORT set to stdio. No account or key: it reads sitemaps that are already public.
Setup effort
One command — uvx sitemap-mcp-server