What people ask, and what we actually answer

No, and no automated tool can. Automated checks find roughly a third of accessibility problems. The rest — whether alt text is meaningful, whether a page makes sense through a screen reader — needs a human being. Scandiary tells you what an automated scan found and what changed since last week. It does not issue a compliance verdict, and no report we send will ever use that word.

The European Accessibility Act is Directive (EU) 2019/882. It has applied across the EU since 28 June 2025 and it covers consumer services including e-commerce. Microenterprises providing services — fewer than 10 staff and annual turnover or balance sheet total of no more than €2 million — are exempt from the service requirements, though that exemption disappears as soon as a business grows past those thresholds. The technical yardstick used in practice is WCAG 2.1 AA, which is what we scan against. We are not lawyers, and this is not legal advice.

WCAG 2.1, levels A and AA. That is the yardstick EN 301 549 points at, and EN 301 549 is the standard European enforcement bodies work from in practice. It is also exactly how the scanner is configured, with the same four rule sets for your site and for ours. We do not check level AAA, and we do not invent rules of our own.

Roughly a third. A machine can tell you that an image carries no alt text. It cannot tell you that the alt text reads image123.jpg, or whether your checkout still makes sense when a screen reader reads it aloud in order. Those parts need a person, and no tool sold to you as a scanner changes that.

No. There is no plugin, no script and no line of code to add. We visit your public pages the way an ordinary visitor would, from our own machines. You give us a starting address and an email address, and that is the entire setup. We never need a login.

You give us one address. We open it, follow the internal links we find there, and keep going until we reach your page limit: five pages on the free plan, twenty on the paid one. We skip files such as PDFs and images, and we skip addresses containing a question mark, because shop filters generate thousands of near-identical pages and would eat the whole budget. In practice that means your front page and the pages closest to it.

Straight away. We scan as soon as you sign up and send that first report the same day, whatever day it is. After that it settles into the weekly rhythm: the scan runs early on Monday and the email follows it.

Yes, and you do not need an account for it. Add each address through the same form, giving the same email address every time, and they are treated as one list: on the Mondays when something has changed, one email arrives covering all of them, with the sites that changed named first and a line at the end counting the ones that did not. The page limit is per site rather than per person — five pages each on the free plan, twenty each on the paid one.

The free plan does not cap how many sites one person watches. The paid plan is bought per site, though, so twenty paid sites would be twenty separate subscriptions at today’s prices — which is a lot, and we know it. If you are watching sites for clients rather than for yourself, write to hello@scandiary.com and say how many, and we will work out a price with you. There is no agency tariff on this page because nobody has asked us for one yet, and we would rather ask you than invent a number.

Two things are worth knowing before you start a long list. We accept twenty new addresses from the same connection in ten minutes, which covers most lists in one sitting; if yours runs longer, wait ten minutes and carry on where you stopped. And an address can be watched by more than one person: if somebody has already asked us to watch it, you are added alongside them rather than turned away, you each get your own copy of the report, and neither of you sees the other’s email address.

Two consequences follow from that, and we would rather write them down than let you find them. The page limit belongs to the address, not to you, so if anyone watching it is on the paid plan, everyone watching it gets the twenty-page scan. And we do not check who owns a site: anyone can ask us to watch any address we can open without signing in, and the report contains only what an ordinary visitor could see anyway. If a report you are expecting never arrives, tell us at hello@scandiary.com and we will say what happened.

No. We see only what an ordinary visitor sees without signing in, and we never ask you for a login. This is a real gap and we would rather name it than let you find it: the checkout is often the least accessible part of a shop, and it is the part we cannot reach.

A dated report written for someone who is not a developer, listing each issue type found, how many elements it affects, and how that compares with the previous scan. If nothing changed, we send nothing — a quiet week is a real result, not a failure. We write “no longer detected” rather than “fixed”, because a problem can disappear because a page was deleted or because an element only appears sometimes. Disappearing is not proof of repair.

No. In a week where nothing differed from the previous scan, nothing is sent: a pointless email is how a subscription gets cancelled. There is one exception and it matters. If the scan could not run, or could not reach your pages, the email still goes out. Silence has to mean one thing only, or it means nothing at all.

Because an issue disappearing is not proof that it was repaired. The page may have been deleted, the element may only appear some of the time, or the scan may not have reached it that week. So we write what we can defend: it was there, and now it is not detected. What that is worth is your call, not ours.

Two things quietly ruin a week-to-week comparison, and both are named before any number is shown: the set of pages we reached changed, or the scan ran from somewhere else. We also scan in a fixed 1440 by 900 window, so that a change in the counts is a change on your site rather than a change in our setup. When we cannot compare honestly, the report says so first and shows the numbers second.

It is not different in what it finds. We use axe-core, the same free open-source engine most tools are built on, so a browser extension will show you the same issues on a page you are looking at. The difference is memory and repetition: we scan the same pages every Monday, keep the history, and tell you what is new, what increased, what decreased and what is no longer detected. Nobody runs an extension on their own site every week for a year, which is the only reason this exists.

No. The scan is axe-core, and the report is assembled from its output by code: the sentences are fixed, and only the findings change. Nothing writes prose about your website. When you reply to a report, a person reads it and a person answers.

We keep the last fifty-two scans of a site, which is about a year. Only the weekly scan in the cloud is allowed to write to that history. A scan run on a laptop can read and compare but never records, because otherwise a test run would land in your record and every later report would carry a warning that had stopped meaning anything.

Five pages every Monday are free, with no account and no card, and that is not a trial that runs out. The paid plan scans twenty pages and lists exactly where every problem is — element by element, not one example — for twelve US dollars a month or a hundred and twenty a year. One email cancels it. Payment is handled by Polar, who are the merchant of record, so your card details never reach us.

They are what the work needs, and they are close to all we hold. This site sets no cookies and runs no analytics, so we genuinely do not know how many people have read this page. Reports are emailed to you and never published. If you want the record removed, write to us and it goes: the privacy page lists every company involved and exactly what each one can see.