Free single-page scan · no account
Find out what your client's page does to disabled users — in under a minute
Paste any page URL. ClearAudit loads the real page in a real browser, runs the same accessibility engine Google Lighthouse uses, and hands you a plain-English list of what is broken, who it hurts, and the exact HTML to fix it. Built for small agencies and online stores who cannot justify a $2,000 manual audit for every project.
How it works
You paste a URL
One page at a time. We accept public http and https addresses, and refuse anything pointing at a private, loopback or cloud-metadata address.
We load the real page
A headless Chromium fetches the page, runs its JavaScript, applies its CSS and lays it out — so what we test is what a visitor gets, not raw HTML.
axe-core checks it, we explain it
The same axe-core engine Google Lighthouse uses runs against the rendered DOM. We group the failures by severity and write up each one for you.
What you get for each failure
- The WCAG success criterion it breaks (for example 1.4.3 Contrast Minimum).
- What is wrong, in plain English — no jargon.
- Who it hurts and how.
- The exact HTML from your page that triggers it.
- Copy-paste code for the fix.
- One sentence telling you what to change.
The offer
Full multi-page agency report — $49
The free scan covers one page. The report runs the same real checks across your site — up to 20 pages, discovered from the links in the pages we render, two clicks deep from the page you submit — and packages the result the way an agency needs to deliver it: the PDF with every issue grouped by type of fix, the WCAG criterion each one breaks, and the copy-paste fix for each one.
- Up to 20 pages crawled, links 2 clicks deep
- Issues grouped by pattern, so one fix clears many pages
- Prioritised fix list, ordered by how much each fix clears
- A branded PDF you can hand to a client
Larger than 20 pages? Then it is sampled, not covered in full: we audit the first 20 pages we reach, and the report says on its cover which limit stopped the crawl and how many pages were never fetched. Nothing is implied about the pages we did not scan.
Our promise: if we can’t produce your report — the site is unreachable, blocks us, or the crawl fails — you get a full refund, no questions.
Have a code? Run the full-site report →Get the $49 report
One field: the address of your site. You get a durable link to your order, and the report is produced once the payment is confirmed. No account, no login.
Full multi-page report
Have a code? Crawl the whole site
Up to 20 same-origin pages, two clicks deep, discovered from the links in the rendered pages. You get a site score, a prioritised fix list ordered by how many pages each fix clears, the issues grouped by pattern, and a branded PDF you can send to the client.
Read a real report before you buy
This is an actual PDF ClearAudit produced — not a mock-up. It is a 14-page report of a public site, made with the same crawler and the same checks the $49 report uses.
Testing sample · public site chosen by us · not a client
www.iana.org — full-site report, 14 pages
Produced 30 September 2026, during our own testing. We picked this site ourselves because it is public and representative. www.iana.org is not a ClearAudit client, has not paid us, has not endorsed us and is not affiliated with us in any way — this is not a testimonial and not a case study. It is simply a real report of a real public site, so you can see the format, the level of detail and the kind of fixes it lists before you spend $49.
- Site score, and a prioritised fix list showing pages affected and pages clean after each fix.
- Every issue with the WCAG criterion it breaks and the copy-paste fix for it.
- An appendix of per-page scores.
- A coverage statement: pages found, attempted and scanned.
Two honest caveats. www.iana.org may have changed since 30 September 2026, so the PDF lists what we found on that day and not necessarily what is on the site now. And it is a sample of the same automated report you would receive — the limits listed below apply to it exactly as they apply to yours.
Open the sample report (PDF, 14 pages)Opens the PDF in a new tab. About 130 KB; nothing else is downloaded.
What this tool is, and is not
It is
- A fast, evidence-based check of the mechanical accessibility failures on a page.
- A way to catch the repeat offenders before a client or a user does.
- A head start on a manual audit — the failures it finds are real, and the fixes are concrete.
It is not
- A WCAG conformance certificate. A clean scan does not make a site accessible, and passing automated checks is not a legal guarantee.
- A replacement for testing with real people who use assistive technology.
- A whole-site audit of every page. The $49 report covers up to 20 pages reached from the links on your site, two clicks deep; a bigger site is sampled and the report says which limit stopped the crawl.
- Legal advice. If you need a conformance claim, get a human expert.
Four things no automated check can judge
Automated checks find mechanical failures — a control with no accessible name, a heading structure that jumps a level, text whose contrast measures too low. These four things need a person, and ClearAudit does not claim to do any of them:
- Whether alt text is accurate. We can prove that an image has no alternative text or an empty one; we cannot tell whether the text that is there actually describes the image.
- Whether the reading order makes sense. We can flag markup that usually breaks it, but only a person moving through the page with a keyboard and a screen reader can say whether the order a user hears is the order that makes sense.
- Whether captions and transcripts are correct. We cannot judge whether captions match the audio, are timed or translated properly, or whether a transcript says what was actually said.
- Whether a task actually works with a screen reader. Signing in, filling in a form, paying at checkout — whether someone can complete the task end to end is tested by people, not by rules.
So: the report is not a WCAG conformance certification and it is not legal advice, and a clean report is not proof that a site is accessible. The $49 report covers up to 20 pages reached from the links on your site, two clicks deep; anything larger is sampled, not covered in full, and the report states on its cover which limit stopped the crawl and how many pages were never fetched.
Privacy — what we keep, and why
Plain version: we keep the minimum needed to run the scan, give you your report and count how many people came — and nothing else.
What we fetch
- The single page you submit to the free scan — that page only.
- For the $49 report: that page and the pages it links to, up to 20 pages, from the same site only, two clicks deep.
- Public http and https pages only.
- The site's robots.txt is respected, and we refuse addresses pointing at private, local, internal or cloud-metadata hosts.
What we store, and where
- Your order — the address you submitted, its status and your order link — in a file on our own server.
- The finished report and its PDF, also on our own server, so your link still works after a restart.
- The email address you leave on the waiting list, in a file on the same server. That address is used only to tell you when the card payment page is working again.
- A count of how many people visit, scan and order, and which link they arrived through. Each page view, each completed free scan and each order adds one line to a file on the same server: the time, the page, and the short tag from the link you clicked if it had one (for example
?ref=linkedin). We count this ourselves, in our own records — no cookie, no third-party analytics and nothing that identifies you. The tag itself is kept in your browser's session storage for the visit only, and — so that reloading the page keeps the same credit — left on the page address you are already on. Both are gone when you close the tab.
What we do not do
- No account and no login — there is nothing to sign up for and no password to lose.
- No third-party analytics, no advertising pixels and no tracking cookies.
- For that counting we store no cookie, no advertising or device identifier, no IP address and no browser user-agent: a counted visit is a time, a page and a tag, and nothing that points at you.
- We never sell, rent or share your data with anyone.
- We do not send marketing email; the waiting list is a single message, sent only to say when the card payment page is working again.
How to have your record deleted
Email [email protected] with the site address you submitted, or your order link, and ask us to delete it. We will remove the order record, the report and the waitlist address, and reply to confirm it is done.
Contact
Questions and refunds
One address for all of it: a question before you buy, a problem with a report we produced, or a refund. If we cannot produce your report — the site is unreachable, blocks us, or the crawl fails — you get a full refund, no questions.