About this automated check
If your website’s logs show a visitor called ClearAisleCheck, that was a small automated accessibility check run by Clear Aisle, a one-person business. This page says what it does, what it never does, and how to ask me not to check your site.
What it is
I fix accessibility problems on online stores: things like images with no text description or buttons with no name, which stop some shoppers from using a page. Before I contact a store, a program opens a few of its public pages and runs the open-source axe-core tests on them, to see whether there is anything I could fix. The result is a report, which I offer to the store for free.
Every request the program makes names itself. Its user-agent line contains “ClearAisleCheck” and the address of this page, so you can always tell these visits apart from shoppers. On 9 and 10 November 2026 the same program called itself “NeevJainSiteCheck”; robots.txt rules written for that name are still obeyed.
What it does on your site
- It reads your robots.txt first and obeys it. That means the rules you set for all visitors and any rules you write for ClearAisleCheck by name. If the file can’t be read, your site is skipped.
- It loads at most three public pages: the home page and, where it can find them, one product page and one collection page. Pages are opened one at a time, a few seconds apart, in an ordinary browser, the way a visitor’s browser would open them. Your visitor statistics may count it as one short visit.
- If it finds problems, it may come back once to read up to three more public pages (the home page, your contact page and, if needed, one page such as “About”), again a few seconds apart. It notes the contact details you have chosen to publish there, so that the report reaches you and not someone else. It doesn’t guess addresses, and it leaves alone any address you have protected against collection.
What it never does
- It never opens your cart or your checkout, never adds anything to a cart and never places an order.
- It never logs in, never opens account pages and never creates an account.
- It never fills in or sends a form, and it doesn’t click buttons or close pop-ups.
- It never tries to get past a block. If your site answers with a bot check, a password page or a refusal, it stops there and leaves.
- It reads only pages that anyone can open without logging in.
How to ask not to be checked
Email me your site’s address and say you don’t want it checked. You don’t need to give a reason. I add your site to a do-not-check list, and after that the program sends your site no requests at all and I don’t contact you.
Email: neevmojang@gmail.com
You can also do it yourself, without writing to me, by adding these two lines to your robots.txt file:
User-agent: ClearAisleCheck
Disallow: /If a report about your store already exists and you want it removed, tell me in the same email and I’ll take it down.
The numbers on the main page
In November 2026 I ran this check on a list of 120 small independent Shopify stores, found through public shop directories and web search. 111 of them could be checked. The others answered with a refusal, a bot check or a redirect, or their robots.txt could not be read, and were left alone.
Across those 111 stores the check read 307 pages: a home page and, where it found them, one product page and one collection page. The figures count stores where the open-source axe-core tests reported at least one failure of that kind on a page that was checked:
- 107 stores had at least one failure of any kind.
- 72 had at least one failure the testing engine rates “critical”.
- 88 had text with too little contrast against its background.
- 64 had links with no readable name.
- 36 had images with no text description.
- 23 had buttons with no name.
This is not a random sample of stores, so it says nothing certain about stores in general. No store is named. Automated checks find only part of what the accessibility standard requires, and a failure reported by the engine can occasionally be wrong.
What happens to what it reads
The report lists the problems found on those pages, with the page addresses, short pieces of the page’s code and screenshots of the parts that failed a check. It is shared with the store it is about, at a private address, and taken down about two weeks later. More detail is on the privacy page.