A WCAG audit that starts from your site’s templates.
We group your public pages into templates and audit one representative page of each against WCAG 2.2.
One page free · no account
We report each measured failure with its evidence and write fix prompts for your coding agent. That is the automated part of a WCAG audit. The rest needs a person testing with assistive technology, which we do not offer.
A scan, an automated audit, an expert audit
They answer different questions. We do the first two. The third finds what the first two cannot.
- A scan
- One page, once. Our free scan checks one public page against WCAG 2.2 Level AA at three screen sizes, and shows one finding with its fix prompt when it can write one. It tells you about that page only.
- An automated audit
- Your templates, on your terms: you set the level, the screen sizes and the pages. Our rules and axe-core’s WCAG-tagged checks measure each page, with evidence for every failed check. A model writes fix prompts and advice, kept apart from measured findings. It cannot judge whether alternative text is accurate, whether a page makes sense read aloud, or how it works with a screen reader.
- An expert audit
- A person tests with a keyboard, a screen reader and zoom, reads the content, and judges what a program cannot. We do not offer it. Our report can give your auditor a list of templates to start from.
Where our audit sits in WCAG-EM’s five steps
W3C’s WCAG Evaluation Methodology describes an evaluation in five steps. We do part of each. How we audit your site says what we record.
- 1. Define the scope of the evaluation
- You choose the site, the level (A, AA or AAA, AA by default) and the screen sizes. We measure against WCAG 2.2 at the level you set.
- 2. Explore the product
- We read your sitemaps and crawl your public links, for up to 2,000 pages or 45 minutes, obeying robots.txt and never signing in.
- 3. Select a representative sample
- We group pages into templates by URL family and bounded HTML signals, not by how they look, and audit a representative page of each. It does not establish what happens on every other page.
- 4. Evaluate the selected sample
- We load each page in a real Chromium browser and measure it at each screen size you set, though some of our checks run only at the narrowest or the widest, and axe-core’s at those two. A check that could not run is reported as not measured, never as a pass. We run no screen reader, so a person finishes this step.
- 5. Report the evaluation findings
- Each measured finding comes with its evidence and, where its rule maps to one, the WCAG criterion. A Coverage tab lists the criteria we could not decide as not tested. We email you a summary when it is ready. We give no score and make no statement of conformance.
What an audit found on our own site
We audited our own site with the product, on our preview server, between 20 August 2026 and 21 August 2026. Here are two of the problems those runs and our own Playwright checks found, as the record states them.
Text areas lose content with WCAG text spacing
- Criterion
- WCAG 2.2 AA 1.4.12 Text Spacing
- Severity
- high
- Where
- The workspace control in the authenticated dashboard, plus one example card on the homepage
- Seen
- Account name clipped in the dashboard workspace control. Content loss in one homepage example card.
- What we changed
- Let the affected text wrap and stop constraining its line box, so applied letter, word, and line spacing cannot clip it.
- Re-audit
- 20 August 2026. Audit 77 re-ran the same target and inspected every eligible rendered text candidate — 51 of 51 at 320, 390, 504, 536 and 604 CSS pixels, 45 of 45 at 834, and 62 of 62 at 1280. No text-spacing finding remained.
Controls were too small to hit
- What the record says
- Found by the first product run on every page and screen size it sampled, and by our own Playwright checks. Audit filters, tabs, selects, the dialog close, back, select-all, audit actions, and the workspace control were all undersized. The shared controls outside the app shell now carry a 44px target. Audit 77 found no undersized control left on the dashboard.
Our audit of our own site has the rest, including the 2 items still open.
Finding, fix, re-audit
An audit is a loop, not a report you file. See what else the product does.
A fix prompt for your coding agent
After an audit we write a fix prompt for each grouped issue, one rule on one page, with its evidence. Copy one into your agent, or build a single prompt from the issues you select, for Claude Code or Cursor or as plain Markdown. Your agent changes your repository. We never do.
Mark it, then audit again
You, or your agent over MCP, mark a task fix applied. A later audit that runs the same check on the page and no longer finds the violation resolves the task and records the date. We confirm that the violation is no longer found, not that the fix is correct. If it comes back, the task is regressed.
On a schedule
A project can run its audit daily, weekly or monthly, on the scope saved with the schedule: the full audit, a template or a URL.
Which WCAG version
We measure against 2.2. Your law or contract may name another.
- WCAG 2.2
- The version we measure against. It contains every criterion of 2.0 and 2.1 except 4.1.1 Parsing, which it removed.
- WCAG 2.1 and 2.0
- A law or a contract may still name one. Check which, and whether it asks you to report 4.1.1. Our guide to what WCAG is sets out the versions.
- Levels A, AA and AAA
- You set the level for the workspace; AA is the default. Our guide to the three levels explains them.
Guides
What a WCAG audit involves, costs and reports, and what to audit first on a shop.
How to do a website accessibility audit in 5 steps (WCAG 2.2)
How much does an accessibility audit cost?
Accessibility audit report: template, example and what to include
WCAG audit for ecommerce: product pages, cart and checkout
WCAG 2.2 AA checklist: every A and AA success criterion, with code fixes
WCAG conformance levels: A, AA and AAA
Questions
What is a WCAG audit?
An evaluation of a site against the Web Content Accessibility Guidelines: you choose a version and a level, test a sample of pages, and report each failure with its criterion. A full audit needs a person with assistive technologies. Ours is the automated part, run on a representative page of each template.
Does an automated WCAG audit make your site compliant?
No. Our audit measures against WCAG 2.2 at the level you set and lists what it found. It gives no score, no certificate and no statement of conformance. A report with no failed checks does not establish that everyone can use your site.
Which WCAG level should you audit against?
The version and level your law, contract or buyer names. Level AA is our default, and you can set A or AAA for a workspace. Our guide to the WCAG conformance levels explains the three.
What does a WCAG audit report include?
Ours lists each measured finding with its evidence, the page and the screen size, and the WCAG criterion where the rule maps to one. A Coverage tab lists the criteria it could not decide as not tested. A report from a person adds who tested and how. See what an audit report should contain.
Do you test with a screen reader?
No. We run no screen reader and test no assistive technology. That belongs to an expert audit.
How is this different from the free scan?
The free scan checks one public page, with no account. An audit covers a representative page of each template, can run on a schedule, and keeps tasks from one audit to the next.
Point us at your site
We scan one page for free and show you what we measured.
One page free · no account