Free website accessibility checker
Paste a page's address and get the real WCAG-oriented checks — images, labels, headings, link text, ARIA, landmarks — plus security headers, SEO basics, speed and broken links. No account, nothing stored.
Free check: overall scores, issue counts and a few example findings. No account needed, five checks a minute, and we don't store your URL or the result.
See a complete sample audit report (print or PDF) → one real, unedited report on a public site — the same document a buyer is emailed, so you can judge it before paying for one.
What this checker runs
Five categories, one page, one pass. These are the checks, in the words the rest of this site uses — the same descriptions as the home page, because they are the same checks. If one can't run, the result says so instead of quietly counting it as a pass.
Accessibility (WCAG-oriented)
Images without alt text, form fields without labels, empty or vague link text, missing or duplicate h1, heading skips, missing lang, unlabelled buttons and iframes, detectable ARIA misuse, missing landmarks.
Security headers
HTTPS in use, http:// → https:// redirect, HSTS, Content-Security-Policy, X-Content-Type-Options, clickjacking protection, Referrer-Policy, Permissions-Policy, and Secure/HttpOnly/SameSite on cookies.
SEO basics
Title and meta description, canonical, mobile viewport, noindex directives, Open Graph tags, robots.txt and sitemap.xml reachability, structured data.
Broken links
A capped sample of the page's own same-origin links is requested for real. 404s and server errors are reported as broken with the HTTP status we got; a link that refuses an automated request (403 or 429) is reported separately as refused, not broken. The report says how many links were checked out of how many were found.
Page weight & speed
HTML transfer size, server response time, an approximate page weight from a sample of the page's assets, resource count, oversized and legacy-format images.
What an automated checker cannot tell you
This is the part that decides whether a report survives being shown to a client. Four things are outside what any markup-level checker — ours included — can see:
Colour contrast
Whether text and background actually contrast enough depends on the rendered colours — inherited styles, images behind text, hover states and opacity. A checker reading the delivered markup cannot see any of that, and neither can ours. It has to be tested on the rendered page, by eye or with a contrast analyser.
Keyboard use
Can every control be reached with Tab, in a sensible order, without a trap, with a visible focus indicator? That is a property of the page as it runs, not of its source, so it needs a person at a keyboard.
Screen-reader behaviour
What a screen reader actually announces depends on the browser, the screen reader and the accessibility tree. Markup-level ARIA mistakes are detectable; the experience of listening to the page is not.
Content quality
Whether the alt text describes the image, whether the link text means something in context, whether the headings make a document outline a person can follow, whether the language is plain — none of that is machine-checkable, and all of it is the part a client notices.
Every report this site produces states all four limits on its own pages, and says which checks could not run. That is deliberate: an agency handing over a document is the one who gets asked “so is it compliant?”, and the honest answer — this is a readiness check with stated limits, and here is what a manual audit would add — is a much better position than a claim the tool cannot support.
The free check vs the $49 client-ready report
Both run the same checks on the same page. The free check is a summary you read on screen; the $49 report is the document you hand over — every finding, its evidence, its fix, and an order to work in.
| In the report | Free check | $49 full report |
|---|---|---|
| Overall score and a score per category | Yes | Yes |
| How many checks ran, and which ones couldn't (with the reason) | Yes | Yes |
| Issue counts by severity and category | Yes | Yes |
| Findings listed individually | 3 examples | Every one of them |
| The evidence for each finding (the element, header or value we saw) | Short excerpt | Full evidence |
| A plain-English fix you can act on or forward to a developer | No | Yes, on every finding |
| A prioritised “fix these first” order | No | Yes |
| How you receive it | On screen, straight away | Emailed to you after you pay, with a private link to the document |
| Ready to print or save as a PDF, with the audited URL and scan date on every page | No | Yes |
The report is a one-off payment for one site. There is no account, no public report page, no monitoring and no white-label option — and the free checker you just ran is the same thing that produces it, so nothing about the findings changes when you pay.
How the check runs
- One fetch of the live page. We request the address you gave us once, read the HTML and the response headers that came back, and score what we received — the page as it is right now, not as it was when a cache was warm.
- Private addresses are refused. Anything that resolves to a local or private address is rejected before a request leaves our server, and again on every redirect hop, so this cannot be used to reach inside a network.
- A 15-second cap per request, a cap on redirect hops, and a cap on how much of the page and how many of its own links we read. A slow or enormous page gets a shorter, clearly-labelled result rather than an open-ended wait.
- We identify ourselves. The request goes out as ProofHiveAuditBot (+https://14b32779c6c2edbf460b6e3805c528e0.ctonew.app; contact proofhive-b69b6585@ctomail.io), so any webmaster whose page we hit can see who asked and write to us.
- Nothing is stored. No account, no database of your URLs, no report archive. The result is returned to your browser and then it is only in your browser.
- 5 checks a minute per visitor, so one person can't take the whole checker. The overall score is a weighted average of the categories that produced one — accessibility 30%, security 25%, SEO 20%, speed 15%, links 10% — and every category score comes only from the checks that actually ran, so a category where most checks were skipped is reported as not scored rather than as a pass.
FAQ
Is this a WCAG audit?
No. This is an automated checker: one pass over the page as delivered to our scanner, running markup-level WCAG-oriented checks — images, labels, headings, link text, ARIA, landmarks — plus security headers, SEO basics, speed and a capped sample of the page's own links. It does not test colour contrast, keyboard use, screen-reader behaviour or content quality, and it is not a certification of WCAG, ADA or anything else. A WCAG audit is a piece of human work; this is the cheap first pass you run before one, or the evidence you use to argue that one is worth commissioning.
What does a WCAG audit cost?
A manual audit is priced as the days it takes, and we could not find a published price for one anywhere — the vendors who sell them quote per organisation. We did collect the published prices we could actually read, from the automated tools and agency suites that compete with us, with a date and a URL for every row. See what an accessibility audit costs — published prices, and the vendors who won't say →
Can I run this on a client's site and send them the result?
Yes — the free check is free to run on any public page, including a client's, and the $49 report is written to be handed over: every finding with its evidence and a plain-English fix, plus a prioritised order. Add your own commentary and charge for the audit as your own work; the report carries the audited site's URL and our name, and there is no white-label option, so say so rather than implying it is your own tool. If you send a client the free check on its own, send it as a summary, not as a compliance statement — it cannot tell you what a manual audit would.
Is a free online checker accurate enough to show a client?
Accurate about what it checks, and silent about the rest — which is exactly the risk. It will not invent findings, and it tells you which checks could not run, but the absence of a contrast finding is not evidence that the contrast is fine, because contrast is never tested here. Free tools differ: WAVE's checker shows a full issue list, the axe browser extension is free, and Pope Tech's free tier covers more pages than this one. Show a client the free check as a summary of what was found, list the things it cannot check next to it (that section is on this page above), and offer the manual audit for those. Do not present any automated checker as proof of compliance — ours included.
Run it on the page, then hand over the document
The free check tells you what is wrong and how much of it there is. The $49 report is that same audit written up as the thing you put in front of a client — every finding, its evidence, a plain-English fix and a prioritised order, emailed to you in full, with a private link that opens it as a document you can read, print or save as a PDF.
See a complete sample audit report (print or PDF) → — one real, unedited report on a public site, with nothing hidden.
Site audit report template — the sections a client-ready report needs, in order, and why each one is there.
Get the full report by email — $49
We email your report within one working day of your payment. If we can't produce your report, you get your money back in full.
These are automated checks on the page as delivered to our scanner at the moment it ran. Accessibility findings come from the markup only — we do not test colour contrast, keyboard use or screen-reader behaviour — so this is a readiness audit and a prioritised to-do list, not legal advice and not a guarantee or certification of WCAG, ADA or any other compliance standard.