Check your own pages against WCAG.
Thirty-one automated checks mapped to specific WCAG success criteria, then a report that says what fixing them actually involves — which findings collapse into a single component change, what your own team can clear, and what order to work in. Results appear immediately: no email, no account. Nothing reaches us unless you choose to send it.
What this does and doesn't cover
We publish an article arguing that automated testing catches roughly a third of WCAG issues, so it would be dishonest to hand you a scanner implying otherwise.
These checks are real and worth fixing — missing labels, unnamed controls, broken heading structure, contrast failures, undersized targets. What no tool can tell you is whether alt text is meaningful, whether focus order preserves meaning, whether captions are accurate, or whether someone can actually complete your checkout. A clean result here is a starting signal, not conformance.
See the full criteria reference for everything a machine can't evaluate.
And then we fix it
The report you get isn't just a list. It groups findings by who actually does the work — content, design, or development — separates the ones where a single component fix clears every instance from the genuine one-offs, and puts them in the order we'd work in.
Download it, print it, or send it over and we'll come back with what clearing it would take. If your team can handle it without us, that's what we'll tell you — we'd rather be the people who said so.
Test the live page, in your own browser
This is the version that actually sees what your visitors see — computed colors, rendered sizes, focus behavior, and content built by JavaScript. A tool that only reads your HTML source can't evaluate any of that.
- Show your browser's bookmarks bar, then drag the button below onto it.
- Visit any page on your site — including pages behind a login, or a staging site.
- Click the bookmark. A report panel opens on the page. Click it again, or press Escape, to close.
Drag that button to your bookmarks bar — clicking it here won't do anything useful, since this page is already ours.
No bookmarks bar? Paste it into the console instead.
Same checks, same results, and it works on sites with a strict Content Security Policy that blocks the bookmarklet.
- Copy the checker with the button below.
- On the page you want to check, open developer tools (F12, or Cmd+Option+I) and pick the Console tab.
- Paste, press Enter. Press Escape to close the report.
Prefer to read it first? View the source — it's about 56KB and does exactly what this page says.
Two honest caveats
- A strict Content Security Policy blocks the bookmarklet. That's good security on the site's part, not a bug — use the console method above, which isn't affected.
- The scan is local. The script runs in your browser against the page you're on, and nothing is transmitted while it runs — we can't see your results. There is one exception, and it is opt-in: if you fill in the form under the report to have us look at it, that submission sends your name, email, the page address, and the findings. Nothing before that.
Paste a page's HTML source
Works anywhere, including on sites that block the bookmarklet. In your browser, view the page source (Ctrl+U or Cmd+Option+U), select all, copy, and paste it here.
Static source only — contrast, target size, and focus behavior need a rendered page, so those checks are skipped in this mode.
Enter a page address
We deliberately don't run a scanning server, and you get better results because of it. Enter the address and we'll copy the checker to your clipboard and open the page — then paste it into the console there.
Why there's no "scan my URL" button here
A page in your browser can't fetch another site directly — browsers block cross-origin requests. So a URL box has to be backed by a server that fetches the page for you, and that server only ever sees the raw HTML the site returns.
That costs you the checks that matter most. Of our 31 checks, 26 can run against raw HTML. The other 5 — colour contrast, target size, focus visibility, and reflow — need a real browser that has laid the page out and resolved its CSS. Contrast alone is the most common accessibility failure on the web, and no server fetch can measure it. A server also can't see anything your site renders with JavaScript.
Running it in your own browser isn't a workaround for not having a backend. It's the better tool, and it happens to also mean nothing about your site reaches us.
You have a list. Now what?
Findings cluster. Twenty missing labels are usually one form component rendered twenty times, and fixing it at the component level closes all twenty at once — which is why we sequence remediation by component rather than by page.
If the list is long, that's normal and it isn't an emergency. Start with anything marked critical: those are the barriers that stop someone completing a task outright.
How Remediation Works