Is there a WCAG 2.2 checker that covers a whole site?

1. Why one page is not enough

Several WCAG criteria cannot be judged from one page at all. Consistent Navigation (3.2.3) asks whether the navigation appears in the same order from page to page. Consistent Identification (3.2.4) asks whether the same control is named the same way everywhere. Consistent Help (3.2.6, new in 2.2) asks whether help is in the same place on every page that offers it. A checker looking at one page has nothing to compare, so it either skips these or guesses.

There is a quieter reason too. Most problems on a site come from a few shared parts: the header, the footer, a card component, a form pattern. One page shows you the symptom; the whole site shows you the cause and how far it reaches.

2. What a whole-site check has to do

Four things, in order.

• Find the pages. From one address, read the sitemap if there is one, or follow the links if there is not, and check that each page still loads. The check should record how it found the pages, because that is part of the scope.
• Check each page. Load it and run the checks for the chosen standard and level, including the states that only appear when someone uses the page: a menu opened, an error message shown.
• Compare pages. Answer the criteria that are about consistency by looking across the set, not one at a time.
• Group by cause. When the same problem comes from one shared component, report it once, with the list of pages it affects, so that one fix and one decision cover all of them.

A tool that does the first two and not the last two has run a page checker many times. That is not the same as checking the site.

3. What WCAG 2.2 changed for the check

WCAG 2.2 kept everything in 2.1 and added nine criteria, six of them at Level A or AA. Two of the six are specifically about more than one page or step: Consistent Help (3.2.6) and Redundant Entry (3.3.7), which asks that a process does not make someone enter the same information twice. The others, such as Focus Not Obscured (2.4.11) and Target Size (2.5.8), apply on every page but are easiest to find across a site, where a sticky header or an undersized icon button repeats. WCAG 2.2 AA in plain words explains each one, and the WCAG 2.2 criteria library has every Level A and AA criterion in one sentence each.

If a contract names WCAG 2.1 or Section 508 or EN 301 549, the check should run against that instead; the right standard is the one the buyer named, not the newest. The Section 508 and EN 301 549 pages explain which WCAG version each one points to.

4. What no checker settles on its own

A whole-site check finds more, and it still cannot decide the criteria that need judgement: whether alt text describes the image, whether link text says where it goes, whether an error message is helpful. Across 25 pages those questions do not get easier; there are more of them. A good tool gathers the evidence for each, puts them in front of a person as a list, and marks them “can’t tell” until that person decides, rather than counting them as passes to make the summary look clean. Why ‘can’t tell’ is an honest answer sets out the reasoning.

So a whole-site check is the start of an audit, not the end of one. The output to look for is not a score across the site but a result for every criterion on every page, with the evidence, and a clear count of what is still waiting for a person.

5. Pages behind a sign-in

Most of what matters on a product sits behind a login, and a checker run from someone else’s servers cannot see it. The honest options are to audit those pages from your own computer, inside your own signed-in session, so that your password never leaves your machine. Dreven offers this two ways: a Chrome extension (added to Chrome by hand) and a command-line tool with an MCP server. Pages recorded or audited that way are reviewed alongside the public ones.

6. What to look for in a whole-site WCAG 2.2 checker

• It finds pages itself, from a sitemap or links, and tells you how.
• It lets you choose the standard: WCAG 2.2, 2.1, Section 508 or EN 301 549, at Level A or AA.
• It checks pages together and answers the cross-page criteria.
• It groups problems by cause, with the pages each affects.
• It keeps evidence for every result and reports four states, with “can’t tell” never counted as a pass.
• It gives a person a place to decide what the tests could not, and it does not claim the site conforms.

7. Questions people ask

Are there WCAG 2.2 checkers that crawl an entire site?

Yes. Dreven takes one address, reads the site’s sitemap or follows its links, checks each page against WCAG 2.2 (or 2.1, Section 508 or EN 301 549) and compares the pages for the criteria about consistency, up to 25 pages at a time. It keeps evidence for every result and sends the results that need judgement to a person rather than guessing.

What is the best tool to check a website against WCAG 2.2?

The best tool for a whole website is one that finds the pages itself, checks them together so that cross-page criteria such as Consistent Navigation and Consistent Help can be answered, groups problems that share one cause, keeps evidence behind every result, and routes the criteria that need human judgement to a person. A tool that only reports a score for one page at a time is a page checker, not a site checker.

Can a WCAG checker test pages behind a login?

Not from the outside. Pages behind a sign-in have to be audited from a browser that is already signed in, on your own computer, so that your password stays with you. Dreven does this with a Chrome extension or a command-line tool with an MCP server.

Does a whole-site check prove a site meets WCAG 2.2?

No. It finds and measures what can be measured on every page in scope and gathers evidence for the rest. The criteria that need judgement still have to be decided by a person, and conformance is claimed page by page, by the people who reviewed the results. A clean automated result is not a conformance claim.

8. Where Dreven fits

Give Dreven one address and it finds your pages, checks them against the standard you choose, compares them for the criteria about consistency, and groups what it finds by cause. Every result carries its evidence and one of four states. What needs a person comes to you as a list, and the report is drafted only from what was measured and what you decided. See what a website accessibility audit covers, or try the free accessibility checker on one public page first.