What the scan sees, and what it cannot

What an ACR is

Universities, hospitals, public agencies and large companies ask software vendors for an Accessibility Conformance Report (ACR) before they sign. Most want it on the VPAT®, the Voluntary Product Accessibility Template that the Information Technology Industry Council (ITI) publishes free at itic.org. The template and its name are ITI's. This desk fills the template in; it is not affiliated with ITI.

The WCAG edition of the template holds one table per level. Each row is one success criterion, a numbered WCAG requirement such as 1.4.3 Contrast (Minimum), and carries a conformance level and a remark. ITI allows four levels: Supports, Partially Supports, Does Not Support and Not Applicable. The remark says why.

How a finding is filed

WCAG 2.2 has 55 success criteria at levels A and AA, 31 at A and 24 at AA, counted from the W3C's published data. The scan runs axe-core 4.14.0, Deque Systems' open-source engine, with 63 of its rules: every one tagged with a level A or AA criterion. Those rules reach 20 criteria. Each violation is filed under the first A or AA criterion its rule names, in axe's own order, so no element is counted twice.

The other 35 criteria have no axe rule at all. Every ledger on this site prints them as Needs a person, at their full count, the sales page included:

1. 1.2.1 Audio-only and Video-only (Prerecorded) A
2. 1.2.3 Audio Description or Media Alternative (Prerecorded) A
3. 1.2.4 Captions (Live) AA
4. 1.2.5 Audio Description (Prerecorded) AA
5. 1.3.2 Meaningful Sequence A
6. 1.3.3 Sensory Characteristics A
7. 1.3.4 Orientation AA
8. 1.4.5 Images of Text AA
9. 1.4.10 Reflow AA
10. 1.4.11 Non-text Contrast AA
11. 1.4.13 Content on Hover or Focus AA
12. 2.1.2 No Keyboard Trap A
13. 2.1.4 Character Key Shortcuts A
14. 2.3.1 Three Flashes or Below Threshold A
15. 2.4.3 Focus Order A
16. 2.4.5 Multiple Ways AA
17. 2.4.6 Headings and Labels AA
18. 2.4.7 Focus Visible AA
19. 2.4.11 Focus Not Obscured (Minimum) AA
20. 2.5.1 Pointer Gestures A
21. 2.5.2 Pointer Cancellation A
22. 2.5.4 Motion Actuation A
23. 2.5.7 Dragging Movements AA
24. 2.5.8 Target Size (Minimum) AA
25. 3.2.1 On Focus A
26. 3.2.2 On Input A
27. 3.2.3 Consistent Navigation AA
28. 3.2.4 Consistent Identification AA
29. 3.2.6 Consistent Help A
30. 3.3.1 Error Identification A
31. 3.3.3 Error Suggestion AA
32. 3.3.4 Error Prevention (Legal, Financial, Data) AA
33. 3.3.7 Redundant Entry A
34. 3.3.8 Accessible Authentication (Minimum) AA
35. 4.1.3 Status Messages AA

Deque's own study of more than 2,000 audits found that automated testing covered 57 percent of accessibility issues. That figure counts issues by volume, and the most common failures, such as low contrast and missing labels, are the easiest to test. Counted by criteria, the scan reaches 20 of 55. Both numbers are true, and a buyer deserves to see the second one.

Why a person sets every level

A rule that finds nothing has passed nothing. axe-core can tell that an image has alt text; it cannot tell whether the text describes the image. It can tell that a field has a label; it cannot tell whether the error message under it says how to fix the error. So in this desk's reports, None found is a note in the remark, and the level beside it is a person's judgement with their name on it.

The reviewer also checks every finding the scan makes. When a rule has misread a page, the remark says so and says why.

Never an overlay

An overlay is a script added to a site that promises to repair its accessibility from the outside. This desk never sells one, never recommends one, and never reports a screen as conforming because one is installed.

On 22 April 2025 the Federal Trade Commission approved a final order requiring accessiBe to pay $1 million. The FTC alleged that accessiBe's claims that its plug-in, accessWidget, could make any website conform to WCAG were false, misleading or unsubstantiated. The order also bars the company from presenting reviews and articles about its products as independent opinions when they are not.

FTC press release, 22 April 2025

That second part is why this site carries no testimonial and no logo wall. One goes up when a buyer agrees to it by name.

What the same report costs elsewhere

Published prices, quoted as each page showed them on the date given:

A lower price can buy a sound report. Ask what it leaves out: whether a person tested the 35 criteria no tool can see, and whether the levels carry a name.

The receipt

Every scan ends in a receipt, a JSON file holding the screens, the time, the axe-core version, the window size, and every finding with its selector and the first 200 characters of the element. It lists the criteria where nothing was found and the criteria that need a person. Its keys are sorted and its SHA-256 is its name, so changing one character changes the name.

When the operator sets this site's signing key, each receipt also carries an HMAC-SHA256 signature that only this server can make and check. Until then every receipt says unsigned: hashed only, and the receipt page says the same.

Anyone holding a receipt can re-run it on the receipt page: the same checks, on the same screens, dated today, with the old and new counts side by side. A report this desk delivers links its receipt in the remarks.

What the free scan cannot see

The free scan opens each screen once, signed out, in one desktop window of 1,280 by 800 pixels, and gives each 25 seconds. It never sees a screen behind a sign-in, a phone layout, a state reached by clicking, or content that loads later. The paid report covers those by hand.

Who signs

The reviewer who signs is not named on this site yet, and no IAAP credential is shown for a buyer to check. An ACR is only as good as the person who signs it, and a desk with no credential on its page is a desk you should ask about. The name goes here the day it is confirmed: [[NAMED REVIEWER NEEDED]].

Sources

• W3C, Web Content Accessibility Guidelines (WCAG) 2.2, Recommendation, 12 December 2024, and its published wcag.json.
• Deque Systems, axe-core 4.14.0, MPL-2.0, rule metadata read from the installed package.
• Deque Systems, Automated testing study identifies 57 percent of digital accessibility issues, 10 March 2021.
• ITI, VPAT®, the template and its editions.
• FTC, FTC Approves Final Order Requiring accessiBe to pay $1 Million, 22 April 2025.
• Every source on this site, with the date it was read, as data: sources.json.