About our scanner

This page exists because our crawler says it does. If you found a request in your server logs from ExhibitA-Scanner or ExhibitA-Research, this explains who we are, what we asked for, and how to stop us.

What the scanner is

Limen is an accessibility auditing tool for online shops, run by one person, Muhammad Talha, in Abu Dhabi. Our server fetches a web page and runs an automated WCAG 2.1 audit against its HTML. The user agent string is:

ExhibitA-Scanner/1.0 (+https://limen.ae/about-our-scanner)

Why we might have fetched your site

For one of four reasons. None of them builds an index, and none follows links off your domain.

• Someone used our free check. Anyone can enter an address at limen.ae/scan, or open a link to it, and our server fetches that one page to check it. We cannot tell whether the person was you. One visitor can start at most five checks a minute and forty an hour, and an identical address is answered from a ten-minute cache instead of being fetched again.
• A customer asked us to check their site across several pages, or to re-check it weekly. We fetch only the address they entered and pages reachable from it.
• We are writing to your business. We sometimes email shops about accessibility problems on their own homepage. Before the email goes, and again at the moment it is sent, we check that homepage so the email describes the page as it is that day. The email links to the free check of the same page.
• We were finding shops to write to. From public company registers (Companies House in the UK, Sirene in France) we look up companies that sell online and read their homepage, and if needed their contact or legal-notice page, once each. These reads check robots.txt first, as ExhibitA-Research/1.0, and stop if it disallows us; the pages themselves are fetched with ordinary browser headers, because many shop platforms refuse unfamiliar clients outright.

If you did not expect a request from us, contact us and we will stop and tell you what we hold.

How we behave

• We obey robots.txt. A Disallow that applies to * or to ExhibitA stops the free check, our customers' scans, our outreach checks and our research reads alike.
• We are slow on purpose. A free check is one page. A customer's scan has at most three requests in flight and a hard cap of 10 to 50 pages depending on their plan.
• Weekly at most. Monitored sites are re-checked no more than once every six days.
• We read, we never write. Every request is a GET. We do not submit forms, follow links that change state, or attempt to log in.
• We give up quickly. Twelve-second timeout and three redirects at most. A single response is capped at 3 MB for customer scans and 6 MB for the free check, because some shop homepages carry several megabytes of inline data.

What we keep

For a free check: the address that was checked, the day, and whether the check came from our own outreach tool or from the public page, for 90 days. Nothing about who started it. The summary of the result sits in Cloudflare's cache for ten minutes and is then gone.

For a customer's one-off multi-page scan: nothing. The audit runs, the result is returned to the customer, and the page content is discarded.

For weekly monitoring we store, per check: the exposure score, the count of findings per rule, and a SHA-256 hash of the page source. We do not store your page source, and we never have. The hash lets us tell whether a page changed without keeping the page.

If we wrote to your business, our privacy notice (in French) says what we hold about it and for how long.

How to block us

Add this to your robots.txt:

User-agent: ExhibitA
Disallow: /

Every kind of request above reads it first, so the next request after you add it is refused, and our outreach will not write to you about a site that has refused us. If you would rather we made no request at all, contact us and we will block your domain at our end.

Contact

Write to hello@limen.ae. We answer within five working days, and sooner if you tell us the crawler is causing a problem.