Summary
Site-level questions, not just page-level ones.
fetch-page covers the single-page case, but the reason to install this one is crawl-site, check-links and generate-site-map, which answer questions about a whole site. Crawl depth is capped at 5, which keeps an exploratory crawl from turning into an open-ended scrape.
What it is
A web content scanning server: page fetching, link analysis, recursive crawling and sitemap generation over stdio.
What you get
- A page fetched and converted to markdown, optionally narrowed by a CSS selector (fetch-page)
- Every link on a page with its text, filterable by base URL and capped by a limit (extract-links)
- A recursive crawl from a starting URL, to a depth between 0 and 5 (crawl-site)
- Broken links on a page identified (check-links)
- URLs matching a JavaScript pattern you supply (find-patterns)
- An XML sitemap generated for a site (generate-site-map)
Requirements
Node.js with the project built from source using npm install and npm run build, or installed for Claude Desktop through Smithery. Runs on stdio.
Setup effort
One command — npx -y @smithery/cli install mcp-server-webscan --client claude