If miromiro-factscan is in your website's logs, that was me. I'm Miro Bzduch, an AI and findability consultant in Wellington, and this is the scanner I run. It reads public pages on New Zealand council and government websites to find facts that a site's own pages state in two different ways, such as a fee, a date or a deadline.

How it identifies itself

Requests carry this name:

miromiro-factscan/1.0 (+https://miromiro.co.nz/contact/; checks public pages for facts that disagree)

If a site turns that name away, the scanner tries once more under a standard browser name, and still follows the site's robots.txt rules for miromiro-factscan. If that's refused too, it stops and leaves the site alone.

What it reads

  • Public pages only. Pages and PDFs anyone can open without signing in.
  • Nothing behind a login, form or security check. It doesn't sign in, fill in forms or run site searches, and it never goes past a challenge page or CAPTCHA.
  • Mostly service pages. Fees, rates, registrations, rubbish, water and similar, ahead of news. It leaves meeting papers and long planning documents alone where it can.

It keeps a copy of each page with the time it was read, so a finding can be traced to what the page said. An AI model helps compare what the pages say, and I check every finding myself before anyone hears about it.

How often, and how gently

  • One request at a time. Requests to a site are 2.5 seconds apart.
  • A limit per site. Up to 400 pages and 20 PDFs in a scan.
  • It follows robots.txt. A rule for miromiro-factscan is obeyed.
  • It isn't a continuous crawler. A site is read in one pass. Later I may reopen a few pages to confirm they still say the same thing.

How to have your site skipped

Tell me through the contact form or at hello@miromiro.co.nz, and I'll stop scanning it and take it off the list. Or add this to your robots.txt:

User-agent: miromiro-factscan
Disallow: /

What happens with what it finds

Findings go only to the organisation they're about. I never name organisations in anything I publish. When I write about a scan, it's totals only.

The scanner supports my work checking what people are told about public services. More on that for councils and in the findability audit.