About the bot

Website Deathmatch lets someone put two websites in a ring and watch them fight over measurable properties. To do that we have to load each site once, in a real headless browser, and write down what happened.

One visit. One page — whatever the URL someone typed resolves to. We do not crawl, we do not follow links, and we never go looking for anything beyond that single page. The browser gives up after 20 seconds.

Before any of that we fetch your robots.txt, and if it tells us not to, we stop there.

Six numbers, and nothing else:

• Page Weight

  Every kilobyte is baggage.

• First Byte

  Hesitate and you get hit.

• Request Count

  A thousand cuts, self-inflicted.

• Domain Age

  Scar tissue counts for something.

• Colour Discipline

  Restraint reads as confidence.

• Tracker Count

  Who else is in the ring with you?

Plus a handful of the colours your stylesheets resolved to, which is what the colour round is scored on. We do not store your HTML, your copy, or screenshots.

At most 5 scans of any one destination per hour, counted against the IP your hostname resolves to — so every subdomain you own shares the same allowance. A measurement is reused for 30 days before we would take another.

There is a second ceiling of 500 scans per 24 hours across the whole site, which holds no matter how the traffic is spread.

We read your robots.txt before every scan and obey it. Disallowing the path — for WebsiteDeathmatchBot or for * — is enough, and there is no setting on our side that skips the check.

If we cannot read your robots.txt at all, we treat that as a no and leave you alone. Only a clear 404 — no rules published — lets a scan through.

You can also block or rate-limit the User-Agent above at your edge. Nothing there needs our cooperation to work, and you do not have to ask us first.