Enter a domain. We read its public pages and give back the pricing, certification, metric and policy claims it publishes — where each one appears, which pairs disagree with each other, which have nothing standing behind them, and which sit on a page nobody has touched inside its own review window.
Reads up to 40 public pages. Obeys robots.txt. Never signs in, never submits a form. What the crawler does.
Every quote in your register is an exact substring of the HTML your server returned, found by 116 published rules — 20 for pricing, 27 for certification, 13 for metric, 30 for policy, 26 for superlative. Nothing is summarised, so nothing can be invented. Read the rules.
The register says what you published, which pairs of your own sentences disagree, which cite nothing, and which sit on pages older than the claim's own review window. Whether a claim is true is your call, and you are the only one who can make it.
If your pages render in the browser rather than on the server, a crawler sees an empty shell — and so would we. When that happens we say the scan failed instead of handing you a clean result. That distinction is most of the point.
Our own homepage published a fabricated price and an uptime figure for a company that does not exist. The write-up, and our own public register.