PropRuleStamp

How to recognise it

Every request carries this User-Agent:

PropRuleStampBot/0.1 (+https://proprulestamp.com/bot)

The scheduled run comes from GitHub's hosted runners, so its IP address varies. Use the User-Agent to recognise it, not an address.

What it does

• It requests a fixed list of public rule, pricing and help pages and help-centre index files (llms.txt). It does not follow links inside pages or discover pages, and it does not fetch robots.txt.
• A scheduled run fetches the list once a day, and some pages once a week. We sometimes start an extra run by hand. Requests go one at a time, at least one second apart; if a page redirects, the redirect is followed straight away.
• It sends plain GET requests. It does not log in, send cookies, submit forms or run JavaScript.
• When a page answers 403 or 429 or shows a bot challenge, it records that, makes no second attempt, asks that site for nothing more in that run, and does not ask for that page again for at least six days. It never retries with other headers, through a proxy or with a headless browser. If a page keeps refusing, we take it off the list; a person may then read it in an ordinary browser.

What we do with the text

We keep the text of each page so that we can compare it with the previous version. When a page changes, we may send its text to a commercial AI service, which proposes new values. A person reviews each proposed change before it is published. The site shows the value with a link back to the page and, for most rules, a short quotation as evidence. We do not republish whole pages.

Opting out

To have a page or a whole site left out of automated fetching, or to ask for a different schedule, write to contact@proprulestamp.com. We will stop fetching it or agree a schedule with you. This does not remove a firm from the comparison.