Free scan

Scan up to 10 pages. See what a machine can find.

Enter a URL, pick how many pages, and we will crawl your site in a real browser and run all 90 supported automated rules: 63 WCAG-mapped checks and 27 best-practice checks reported separately. Each automated finding names the source rule, with a direct fix guide when our library has a verified match. It runs safely in a queue, emails you when it is ready, and stays honest about its limits. We graded 356 of the W3C's 432 techniques and documented failures for what a machine can settle, and by that grading of ours a machine fully settles 10 of them. The rest is covered by the human audit.

Choose what to scan

You can leave this page after submitting. The worker holds your place, your dashboard shows progress, and we email you when the result is ready.

A free account holds your queue place.

Enter the website below first if you like. We carry it through registration or login, then bring you back here.

Create account Log in

What it checks

What this configuration of the scanner can see
Missing alt attributes, unlabelled fields, contrast below the threshold, ids that break an ARIA reference, empty links and buttons, and the rest of what these rules cover. Not every machine-detectable failure there is, because no single rule set is that.
Ten pages, not one
A single page tells you about a template. Ten tells you whether the problem repeats, which decides whether this is an afternoon or a project.
Each finding names its source rule
Every result names the automated rule that produced it. WCAG findings link to their criterion, best-practice findings stay separate, and a direct fix guide appears wherever our library has a verified match.
Which findings need a person
Results are marked by whether a scanner settles them or a human has to judge, so you can see the shape of what is left.

What it cannot tell you

The 90 supported automated rules include 63 WCAG-mapped checks and 27 additional best-practice checks, which the report keeps separate. The WCAG-mapped checks touch 20 of the 55 Level A and AA criteria we audit, and touching is not the same as clearing. A clean result says nothing about the other 35 in either direction.
Whether your alt text is correct. “A dog” on a photograph of a cat passes every automated check there is.
Whether a journey can be finished. A crawler visits addresses one at a time and cannot put something in a cart and carry on, and conformance across a process needs every step in that process to conform, so the question you most want answered about your checkout is the one this cannot touch.
Anything behind a login, and any state you have to do something to reach. Each page is checked once, in the state it arrives in, so the scan opens no menus, submits no forms, triggers no error messages and sits out no timeouts. The error nobody announces only exists after somebody presses the button.
How your site behaves on a phone. Every page is rendered at 1366 by 900 and at no other width, and the standard treats each layout a responsive page presents as a variation that has to conform on its own, so this is a missing half of the evidence rather than a detail.
Anything at an address it will not accept. The scanner takes a public hostname on a standard web port and refuses everything else, so localhost, internal hostnames, private networks and staging on an odd port cannot be scanned at all. And a site with fewer than ten crawlable pages returns fewer than ten.

Where to go next

What this scan finds

Machine-facts: images with no alt attribute, form fields with no label, contrast that measures short, missing page language, broken ARIA references, empty links and buttons. Every finding names its source rule; WCAG and best-practice results stay separate, with a direct fix guide wherever our library has a verified match.

What machines can and cannot check

What it cannot find

Whether the alt text actually describes the image. Whether keyboard focus gets trapped in your menu. What a screen reader announces at checkout. Whether error messages make sense. That takes a trained reviewer and a real blind screen-reader user: the other two passes of our audit.

Why the ceiling sits where it does

None of that is our scanner being modest. W3C publishes 90 test rules for WCAG, 87 of them live, and each one states what its own result means. On 59 of them a clean pass is recorded as leaving the criterion still needing further testing. Exactly four can clear a criterion outright. And 49 of WCAG's 86 criteria have no published test rule at all, so more than half the standard sits outside the reach of any conformance-tested automated rule, and that is settled before anybody starts arguing about which scanner is best. Lighthouse prints its own version of this warning at the top of every accessibility report, and WAVE's help says outright that only a human can determine true accessibility. Which rules a scan can and cannot settle is set out one at a time.

So run it, and treat each result as evidence to investigate, not a verdict by itself. A located automated finding is worth checking, and the report gives you the exact page, context, element, rule, and explanation to do that.