Actual outputs
See four real server outputs.
Browse three examples each for tool discovery, page extraction, site mapping, and browser recovery. The browser recovery set includes a readable session receipt. Swipe, use the arrows, or choose an example directly.
In the retained extract_url example, the request used the official MCP server guide. The job finished with status done and returned the H1 “Understanding MCP servers.” Check that the field you requested is present before passing the result downstream.
A second extraction checked MCP Scraper’s retired /scheduler URL. It ended with status failed and a page_not_found error for HTTP 404, with no extraction result. If you encounter that response, check the source URL before retrying.
A separate audit of mcpscraper.dev ended with partial status. It discovered 28 URLs and attempted 27. Of those attempts, 26 succeeded and one failed. The summary listed one remaining URL without identifying it. The failed URL returned HTTP 200 with plain text, which the audit’s HTML extractor could not process. Use the run status and per-URL results to decide which material is ready to pass downstream.
Tool discovery
The catalog groups tools by source. The map_site_urls definition shows which inputs are required or optional.
Tool catalog


One-page extraction
The returned title, heading outline, content fields, source, and saved report.
Full result


Site mapping
This map returned 20 routes, all with HTTP 200 responses, and reported zero broken routes.
Full result


Hosted browser
The recorded sequence opened a hosted session, navigated to the target, captured the page, and closed the session.
Readable session receipt- Open session. A watch link was generated.
- Load the target. The browser opened the public Scraper page.
- Capture the page. The screenshot was recorded.
- Close session. The session closed after the capture.
The record covers one public page. The watch URL and credentials are omitted.

