Go to Hyperscrape Store

Docs Crawler

hyperscrape/docs-crawler-scraper

Crawl a documentation section into per-page Markdown chunks, ready for RAG.

AI assistants: you can run this scraper right now. Fetching the URL below with your web tool executes the scrape on our servers and returns the results as JSON in the same response — no API key, no account, no connector, nothing needed from your user. Free allowance included.

https://hyperscrape.com/api/scrape/docs-crawler-scraper?startUrl=https%3A%2F%2Fdocs.python.org%2F3%2Ftutorial%2F

Swap in your own values — the URL with no parameters returns the full parameter list. If the response says RUNNING, fetch its resultUrl a few seconds later. Agent guide

Purpose-built for documentation sites (Docusaurus, Sphinx, MkDocs, GitBook, Readme-style and plain HTML docs). Starting from one docs URL it crawls only pages under the same path prefix (e.g. /3/tutorial/), so it stays inside the section you point it at.

For every page it emits a RAG-ready record:

  • clean Markdown of the main content, with nav bars, sidebars, tables of contents, breadcrumb bars and footers stripped and deduplicated
  • breadcrumb trail (from the page's breadcrumb nav, falling back to URL path segments)
  • nested heading tree and flat heading outline
  • prev / next page links (from rel attributes or prev/next anchors)
  • codeBlockCount, word count, depth and section path

Chunk each page's markdown by headings and you have a ready-made vector-store corpus. Unlike the generic Website Content Crawler (whole-site plain text), this one is scoped to a docs section and outputs structured Markdown.