Give us a URL.We read it as they do.

We fetch your pages without running a line of your JavaScript, which is what every AI crawler does. Nothing to install and no account.

Up to twenty-five pages

Taken across the whole sitemap rather than the first few, because the first few are the ones somebody has already hand-tuned.

Four at a time, and we say who we are

The user-agent is almost.sh-scanner. We never impersonate GPTBot or anybody else: not running your JavaScript is what reproduces what a crawler sees, not the name on the request.

Nothing is stored but the findings

The report, and a hash of the address it was requested from so the same person cannot queue a thousand. No page content is kept.