Kept Record

KeptRecordBot visited your site. This is what it is.

What it is

KeptRecordBot is the web browser of Kept Record, a service that measures accessibility defects on storefronts and keeps a record of them over time. It is a real Chromium browser, driven by a program, that loads a page the way a visitor would and then reads the page's structure. It does not index, cache or republish your pages, and it does not train anything on them.

Kept Record is an independent service, run by its founder and reachable at idan@keptrecord.co. If you have a question or an objection, write to idan@keptrecord.co.

Why it visited

One of two reasons. Either someone entered your page's address into the free scanner at keptrecord.co, in which case it read that one page, once, at their request, and nothing else on your site. Or the owner of your site, who first proved control of the domain by publishing a DNS record we gave them, asked us to scan it, in which case it read up to a set number of pages, following links within your site only, on a schedule the owner chose, at most once a week, or when the owner asked.

Nobody can ask us to crawl a site they have not proved they control. The free scan reads one page and follows no links.

How it identifies itself

Every page request carries this user agent, so you can find it in your logs:

Mozilla/5.0 (compatible; KeptRecordBot/0.1; +https://www.keptrecord.co/bot) accessibility-defect-scanner

Where it comes from

Every visit comes from one address, 2.28.228.80, a server in Nuremberg, Germany. Allow it or refuse it by that address if you prefer that to robots.txt. Before a scan, the service only resolves your host name in DNS; it does not fetch your page from anywhere else.

Before reading any page it fetches your robots.txt, and that one request identifies itself as KeptRecordBot alone.

How to refuse it

Add this to your robots.txt and it will not read any page on your site. It re-reads robots.txt at the start of every scan, so the change takes effect on the next visit.

User-agent: KeptRecordBot
Disallow: /

What it obeys in robots.txt

  • Disallow rules addressed to KeptRecordBot, and rules addressed to every bot with an asterisk.
  • Crawl-delay, which can make it slower than its own setting and never faster.
  • If your site has no robots.txt, or the file cannot be fetched, it reads the page. It does not treat a missing file as a refusal.
  • The owner of a site, having proved control of the domain, can switch the robots.txt check off for that one site. Nobody can switch it off for a site they do not control, and the free scan always obeys it.

How it behaves on your pages

  • It reads. It fills in no form, submits nothing, clicks nothing, and signs in to nothing.
  • It runs your page's own scripts, as any browser does, so your analytics and any third-party code on the page will see a visit.
  • It starts every scan with an empty browser profile and discards it afterwards. It keeps no cookies between visits.
  • It waits at least one second between page requests, longer if your robots.txt asks, and reads at most five megabytes of any one response.
  • If your site answers with a refusal, such as a 403 or 429, or a challenge page, it records that and stops. It does not retry with a different identity or from a different address.

What it keeps

For each page: the address, the title, the HTTP status and how long it took to render. For each accessibility defect: the failing element's markup and its accessible name, after email addresses, phone numbers, long runs of digits and every form value in them have been replaced with placeholders. It stores no screenshots and no full pages.

Records from the free scanner are deleted 7 days after they were requested. Records for a site whose owner asked for scans are kept for that owner, for as long as they keep their account.