Accessibility audits by real people
Can everyone finish what they came to do?
An expert checks all 55 WCAG 2.2 A and AA rules. A blind screen-reader user works your real journeys. You get the barriers, the evidence and the fixes.
Rapid Audit · $499 · up to 10 pages or journeys
Your web report and PDF in 5 business days from cleared payment.
See what you get: read the sample reportWhat every audit promises
No overlays
Read why we skip them
Real evidence
Every finding has proof
Fixed price
$499, no surprises
Fixed turnaround
5 business days
Your next step
What do you need to do today?
Free · account required
Check a website
Scan up to 10 public pages and get automated findings linked to fix guides. Human testing comes next.
Run the free scanFree · no account needed
Fix a known issue
Describe what is going wrong. Find the WCAG rule behind it, then follow its practical fix guidance.
Find the ruleAudits from $499
Plan a human audit
Compare coverage and turnaround. Get expert review, blind screen-reader testing and a report your team can act on.
Compare audit options
Why humans
A clean scan does not mean your site works
In accessibility, no news is not good news. W3C publishes 87 live test rules for WCAG, and those are the checks a scanner automates. What a clean pass actually settles is recorded in each rule's own outcome mapping, and we counted the whole set.
What the 87 test rules can settle
59
record a clean pass as “needs further testing”
24
name no WCAG criterion at all
4
can clear a criterion outright
Underneath that sits a wider gap. Of the standard's 86 criteria, 49 have no test rule written against them at all.
That means the people who write the standard are telling you, on rule page after rule page, that passing the automated test is not the same as meeting the rule. So a failure from a scanner is worth acting on the same day it appears. A pass is not an answer. It is a question the standards body has already told you is still open.
Try it yourself
The thirty-second checkout test
Unplug your mouse, open your own checkout, and tab your way to the button that takes the money. Now listen to what it announces. If all it says is “button”, a scanner will still call that page clean, because the button exists and it has a name of sorts. A spell checker has the same blind spot. It can tell you that teh is misspelled, and it cannot tell you the sentence makes no sense.
A keyboard flow that traps focus raises no warning either, and neither does a checkout that cannot be finished by ear, or alt text that confidently describes the wrong photo. Those surface when a person works your pages with the software they actually use, which is why every WCAGrules audit includes a hands-on session with a professional blind screen-reader user, and why we never install an overlay widget, which does repair a real set of problems as the page loads and leaves your code exactly as it was. The full split of what machines can and cannot verify is public in our scanner-versus-human breakdown. So that is the argument for doing it by hand. Below is how those five days actually run.
How it works
Five business days, counted from the day payment clears
Before the clock starts
You pick the pages, we confirm the scope
Choose up to 10 pages or journeys, or hand us the choice and we will take the ones your money moves through. You pay as you order, on Stripe's own checkout page, and the five days start the moment payment clears. We confirm the exact page list by email the same day.
Days 1 to 2
The scan runs, then a person checks the rest
The machine goes first, because it is quick and it settles what a machine can settle on its own. Then an auditor works every page against all 55 WCAG 2.2 A and AA rules by hand, which is where the judgment calls get made.
Days 2 to 4
A blind auditor works your real journeys
A professional screen-reader user takes on the real tasks, finding a product and getting through checkout. The failures a scan cannot reach get found by somebody who hits them, and every one is screenshotted the moment it happens.
Day 5
Your report lands with the evidence attached
A web report and a PDF, both yours. Every finding carries the rule it breaks, its severity, a screenshot, and the fix that closes it, so the work can start that afternoon.
The honest limit, before you buy. An audit covers the pages in your scope and nothing else. W3C's own evaluation methodology says plainly that testing a sample of a site cannot carry a conformance claim for the whole site. So this report tells you what is broken on the pages that matter most, and it is not a certificate for the other ten thousand. Want to see what those 55 rules actually ask of a page? Read all 55 rules, one page each.
Learn the rules
All 55 rules. Checked, every time.
WCAG 2.2 sorts its success criteria under four principles, and your report walks all four. Every criterion that applies to your pages comes back as a pass, a fail, or a not applicable, with the evidence sitting behind it. Every rule below links to its own fix guide, free, whether you ever buy anything or not.
The three rules to start with
Browse all 55 rules, explained and searchable, with examples, fixes, and what each one looks like when it fails an audit. Or see what the report looks like.
The report
A work plan with proof attached
Each failure gets named, along with who it shuts out, the proof behind it, and the fix that closes it. Findings arrive ordered by user impact, so nobody has to guess where to start. It is fair to say that a findings list your team cannot reproduce is just a PDF, which is why every line in ours carries the screenshot and the exact element it was taken from. A scanner would've handed you a score.
- Every finding tied to the WCAG 2.2 rule it breaks, and graded for severity
- A screenshot of each problem, taken the moment a real user hit it
- What the problem blocks, written so anyone on your team can read it
- The fix, drawn from our free library of 432 guides and picked for your stack
- A work plan ordered by user impact, so your developers know what to do first
Product images have no text alternative
On 6 of 10 audited pages, product photos are announced only as a file name, so a screen-reader user hears your catalogue as a list of camera settings and cannot tell one product from another.
How to fix
Give each product image alt text that names the product, so alt="Gold twist ring, size 7"
Evidence: screenshot · the exact element · the page state it was found in
Who it is for
You own the website. This is your audit.
That report has two jobs. It shows you exactly what is broken, and it hands your team everything they need to fix it. Here is one finding, read both ways.
Product images have no text alternative
Screenshot captured 2026-08-22 · exact element pinned
How you read the finding
Your product photos read as file names. A customer using a screen reader cannot tell one product from another.
- What broke, who it shuts out, and how bad it is, in plain words
- A dated screenshot pinned to the exact element that failed
- Findings ordered by user impact, so the worst blockers come first
How your developers read the finding
Reproduce
NVDA reads the photo as “graphic, IMG_2041.jpg”.
Fix
Add alt text naming the product. One attribute.
- The exact element, the rule it breaks, and the steps to reproduce it
- Guidance picked for your stack, so nobody has to translate it first
- Each finding linked to its fix guide, from the free library of 432
The finding above is an illustration, drawn in the format your own findings arrive in, from a failure we see constantly rather than from a client engagement.
Pricing
Fixed prices, no discovery calls
Free Scan
$0
Up to 10 pages · automated only
For the owner who wants a first honest look before spending anything.
- Every page loaded in a real browser
- Findings mapped to the rule they break and the guide that fixes them
- A straight count of what still needs a person
- Covers what a machine can prove, which reaches 20 of the 55 rules
Rapid Audit
$499
Up to 10 pages or journeys · 5 business days
For the owner who wants the truth about their most important pages this week, without a discovery call first.
- You pick the pages, so home, category, product and checkout
- Automated scan of every page
- Expert review against all 55 WCAG 2.2 A and AA rules, a pass or fail recorded for each
- Hands-on session with a blind screen-reader user (NVDA or VoiceOver)
- Findings ordered by impact, with screenshots and fixes
- Web report and PDF, evidence attached to every finding
- Half-price re-audit within 3 months
Standard Audit
$1,499
Up to 25 pages · 10 business days
For the owner who needs the whole journey covered, page by page, with both screen readers on it.
- Everything in Rapid
- Extended screen-reader sessions (NVDA and VoiceOver)
- A table recording pass or fail against all 55, for every page tested
- 25 pages, so the coverage reaches across your key journeys
Agencies and dev teams. Audit a client's site before launch and hand the report straight into the build, at the same fixed price with the same evidence. Email hello@wcagrules.com for volume pricing, or see the full pricing page.
Free to read
Use the research behind the audit
Learn the rules, look up a fix or check a requirement before you spend anything. The library explains the standard. A scan and a human audit help you apply it to your own pages.
FAQ
What owners ask before they order
What do I actually get for $499?
A hand-run WCAG 2.2 AA audit of up to 10 pages or journeys, in three passes. A scan first, then an expert review of every page against all 55 Level A and AA rules, then a working session with a professional blind screen-reader user on NVDA or VoiceOver. The web report and the PDF land inside 5 business days of payment clearing, and every finding in them carries its evidence and its fix, which is what lets your team start on the Monday rather than start arguing.
Who does the testing?
A trained accessibility auditor first, then a professional blind screen-reader user who works in NVDA and VoiceOver every day. A Rapid audit gets a session on whichever of the two suits your audience, and Standard runs extended sessions across both. Testing is desktop-based, and everything the tester hits goes into the report as a screenshot rather than as a description you have to take on trust.
How much of this can I read before I spend anything?
All of it, and we mean all of it. Every one of the 55 A and AA rules is explained on this site, along with the 31 AAA ones, 432 fix guides, 17 free tools, 38 accessibility laws, and the full sample report. The free scan covers up to 10 pages with a free account. The honest limit is that reading the library tells you what the rules say, not which ones your site is failing, and only somebody working your pages answers that.
Can my own developers fix what you find?
That is the whole point of the report. Every finding names the exact element, shows the screenshot, and links a step-by-step guide from our free library of 432. Most teams start fixing the day it lands, because nothing needs translating first. There is no retainer and no locked portal, and we would rather your own team owned the fix anyway, since they are the ones who have to live with the code.
Do you sell the repair work too?
No, and that is deliberate. We find the problems and prove them. Your team fixes them, or your agency does, and we verify the fixes on a re-audit at half price inside 3 months. The truth is, a firm that gets paid per finding and paid again to fix them has a reason to find more, and we would rather not have that reason sitting in the room while somebody counts your problems.
Do you install an accessibility overlay or widget?
Never. An overlay does repair a real set of problems as the page loads, and it leaves the code underneath exactly as it was, so the defects a complaint would cite are still sitting there for anyone to find. The practitioner reference on overlays is explicit that full conformance cannot be reached with one. We go after the root causes instead, so your team closes them for good and stops paying rent on the symptom.
Is my report private?
Yes. We publish nothing about your site. The report is yours to use internally, hand to your developers, or show a lawyer, a court, or a regulator. The one thing it is not is a badge, so it cannot be republished as marketing. A findings report says what we found, and it never says we endorse you.
How do I pay?
You pay as you place the order, by card, on Stripe's own checkout page, so no card details ever touch this site. We confirm the exact page list by email the same day, and if the scope turns out not to fit, we refund in full before any work starts. The 5-day clock runs from the moment payment clears.
Find out where you stand. Five days, one price.
A fixed $499 for up to 10 pages or journeys. Expert review, blind screen-reader testing, and evidence your team can use. Delivered in 5 business days from cleared payment.