Patrick Stox Tools Fetcher

Some tools fetch a URL only after a visitor asks them to check it. This page explains exactly how those requests identify themselves and how to stop them.

Our user agent

Requests made by the standard, non-emulation tool fetcher use:

PatrickStoxTools/1.0 (+https://patrickstox.com/fetcher/)

What we fetch

A fetch happens only when someone submits a URL or host to a tool that needs a live response, such as a status, redirect, robots.txt, sitemap, or HTML check. We do not run a general web crawler, discover pages on our own, or maintain an index of fetched pages.

A few diagnostics deliberately emulate a browser or a crawler selected in the tool (for example, user-agent comparison checks and mobile rendering). Those requests use the displayed simulated identity rather than the standard string above, because that is the test being requested; they are still on-demand checks, not autonomous crawling.

Requests are bounded by URL validation, redirect limits, size limits, timeouts, per-host pacing, and abuse controls. Tool input and fetched page content are not retained as a crawl archive.

How to verify a request

For a standard fetch, check the complete user-agent string above and the request behavior: it is a small number of on-demand GET requests following a visitor action. For an emulation check, use the identity selected by that tool instead. Fetches run from Cloudflare infrastructure, so egress IP addresses can change and do not have a stable forward-confirmed reverse-DNS identity. Do not rely on IP allowlists or reverse DNS to identify this fetcher.

How to block it

If you do not want these user-triggered checks, block the exact user agent at your CDN, WAF, or origin. A robots.txt rule can communicate a crawler preference, but it is not an access-control mechanism for a visitor-initiated diagnostic request. Blocking the user agent at the server layer is the reliable option; the person using the tool will see that the page could not be fetched.

If you believe a request using this user agent is abusive, please include the timestamp, target URL, and relevant request headers in a contact request.