Free accessibility checker
find the barriers excluding your customers
Scan your page for the common technical failures cited in ADA complaints — classified into what is confirmed, what is uncertain, and what needs human review. The legal exposure is real (3,117 federal suits in 2025, most against small businesses), but the person locked out of your checkout is the reason to fix it.
Reads one page the way a search engine would. We do not store your page or submit forms.
Who is getting sued, and for what
Federal web accessibility filings reached 3,117 in 2025, up 27% on 2024 and the second-highest year on record. Counting state courts, the total passed 5,000. Web cases were 36% of all ADA Title III federal filings, up from 28% the year before. New York, Florida and California still dominate, and Illinois went from 28 cases to 237 in a single year.
The targets are mostly not national brands — UsableNet's tracking puts 64% of the companies sued in the first half of 2025 under $25M in annual revenue. This is a volume practice, and the usual first contact is typically a demand letter seeking a settlement — commonly five figures, per defence-firm reporting — to avoid a filing. That is what makes a cheap, honest scan worth running: the issues being cited are, overwhelmingly, the ordinary ones.
Sources: Seyfarth ADA Title III, EcomBack 2025 annual report, UsableNet 2025 midyear report.
Three kinds of finding, never mixed
Most scanners return one undifferentiated list of “errors”. That is how you end up spending a morning fixing something that was never broken, and then distrusting the items that were. Every finding here declares what kind of claim it is making.
| Category | What it means | Counts toward your score? |
|---|---|---|
| Confirmed failure | The evidence in your page proves it. An alt attribute is either absent or it is not; a contrast ratio measured from two known colours is arithmetic. | Yes, at full weight |
| Potential issue | A real signal, but the criterion allows for context the page does not contain. “Read more” may be fine if the card title beside it does the work — the criterion is Link Purpose in Context. | Yes, at reduced weight |
| Manual review required | We can see that something needs deciding and we cannot decide it. Whether a video has speech. Whether your alt text is accurate. Whether a table holds data or is only there for layout. | No. A score implies we know, and here we do not. |
The test applied to every check is the same one: what did we actually observe, and what are we assuming on top of it?“This video element has no caption track” is an observation. “This video needs captions” is an assumption about audio we cannot hear. The first is reportable; the second is a question for you.
We found that mistake in our own engine and it is why this exists: four silent screen recordings on our own homepage were reported as a critical captions breach. There was nothing to caption. That finding is now a manual-review item and scores zero.
What this checker tests for
All 14 checks, in the order the scanner reports them — which is the order they appear in filed complaints, not the order they appear in the WCAG document. 1 always produce a corrected value you can paste in and 2 sometimes can. The rest locate the problem and leave the wording to you, because there is no correct value a machine could derive from the page.
| Check | WCAG criterion | What kind of claim | Confidence | Value computed |
|---|---|---|---|---|
| Images with no alt text | 1.1.1 Non-text Content (A) | Confirmed failure | High | No |
| Buttons and links with no readable name | 4.1.2 Name, Role, Value (A) | Confirmed failure | High | No |
| Form fields with no label | 3.3.2 Labels or Instructions (A) · 1.3.1 Info and Relationships (A) | Confirmed failure | High | No |
| Embedded frames with no accessible name | 4.1.2 Name, Role, Value (A) | Confirmed failure | High | Sometimes |
| Page language not declared | 3.1.1 Language of Page (A) | Confirmed failure | High | No |
| Missing or empty page title | 2.4.2 Page Titled (A) | Confirmed failure | High | No |
| Text contrast below the WCAG minimum | 1.4.3 Contrast Minimum (AA) | Potential issue | Medium | Sometimes |
| Link text that may not describe its destination | 2.4.4 Link Purpose in Context (A) | Potential issue | Medium | No |
| No quick way to skip past the navigation | 2.4.1 Bypass Blocks (A) | Potential issue | Medium | No |
| Heading levels that skip a step | 1.3.1 Info and Relationships (A) | Potential issue | Medium | No |
| Positive tabindex forcing a custom focus order | 2.4.3 Focus Order (A) | Potential issue | Medium | Yes |
| Table with no header cells | 1.3.1 Info and Relationships (A) | Potential issue | Medium | No |
| Video or audio with no captions or text alternative | 1.2.2 Captions (A) / 1.2.1 Video-only (A) | Manual review required | Unknown | No |
| Media set to play automatically | 1.4.2 Audio Control (A) | Manual review required | Unknown | No |
- Embedded frames with no accessible name — sometimes computable. Only for embeds whose purpose is unambiguous from the URL - a YouTube address is a video player. For any other host we say we cannot name it, rather than describing content we have not seen.
- Text contrast below the WCAG minimum — sometimes computable. A passing colour is computed unless no lighter or darker version of your text colour clears 4.5:1 on that background, in which case the background is the thing that has to change.
- Positive tabindex forcing a custom focus order — always computable. tabindex="0" is always the correct replacement for a positive value.
What no automated scan can find — including this one
Automated tooling detects somewhere around 30 to 40 percent of what WCAG actually asks for. The rest needs a person, and the following are the specific things this scanner does not and cannot judge. They are shown with every report, not buried in a footnote, because a clean automated result on a page with a keyboard trap is worse than no result at all.
- Keyboard traps — whether every interactive element can be reached and left using only a keyboard
- Focus visibility — whether the focused element is always clearly indicated on screen
- Screen reader behaviour in custom widgets such as modals, carousels, date pickers and menus
- Caption and transcript accuracy — a captions file can exist and still be unusable
- Whether alt text is actually correct, rather than merely present
- Reading and navigation order once CSS has positioned the page
- Colour contrast on text whose colours come from external stylesheets or JavaScript
- Form error handling — whether mistakes are announced and explained to assistive tech
- Whether a heading outline is logically organised, as opposed to merely unbroken
- Whether a table presents data or is only there for layout
- Whether generic link text is rescued by the sentence or card it sits in
Overlay widgets, the accessiBe FTC order, and why this page will never tell you that you are compliant
Every other scanner in this category leads with some version of the same promise. The biggest one, accessiBe, headlined its free scan with checking whether your site is “accessible and ADA compliant in seconds”. In January 2025 the FTC ordered the company to pay $1,000,000, alleging its automated widget did not make websites WCAG-conformant as advertised and that it had presented paid reviews as independent endorsements. The final order bars it from claiming its automated products can make any site WCAG-conformant without evidence to back it up.
That is the whole reason this tool talks the way it does. A scanner that tells you the answer you want is worth less than one that tells you what it actually saw — because the false reassurance is what leaves the real barrier in place, and the real barrier is what gets cited. We report findings. We name our limits. We do not sell an overlay, and we would advise you not to buy one.
Source: FTC press release, January 2025.
Six questions to ask any accessibility checker
Including this one. Every answer below is checkable on the page you are reading, which is rather the point.
| Question | This checker | Typical in this category |
|---|---|---|
| Does it tell you the check list? | Yes — all 14 are published below, with the WCAG criterion each maps to. | Usually a headline number (200, 320, 396) with no list. A "check" is often the same rule counted once per element. |
| Does it claim to make you compliant? | No. It reports what it found and states what it could not see. | Frequently yes. The FTC fined the largest vendor $1M in 2025 over exactly that claim. |
| Does it say what it cannot test? | Yes — 11 named limits, listed below, shown with every report. | Rarely stated, and almost never on the results screen where it matters. |
| Does it give you the corrected value? | Where it can be computed, yes — a contrast-safe colour verified against your actual background, tabindex corrections, embed titles. Not alt text: correct alt text depends on what the image is doing on the page, and we will not invent it. | Commonly a rule name and a link to the WCAG spec. |
| Does it need a signup to see results? | No. Paste a URL, read the findings. | Often an email gate before the report, because the report is the lead magnet. |
| Is it trying to sell you an overlay? | No. We do not sell one, and we think you should not buy one. | Often the entire business model behind the free scan. |
How the report is built
Ordered by what actually gets sued
Findings are ranked by how often each issue is cited in complaints, not by WCAG chapter. Missing alt text, unlabelled form fields, unnamed controls and uncaptioned video come first because that is what the complaints are built on.
Where that ranking comes from, so you can check it: we did not code court filings ourselves. The ordering follows the published litigation trackers — UsableNet’s and EcomBack’s annual reports, which categorise the barriers alleged in the complaints they track — and those reports consistently put the same handful of technical failures at the top. Where the trackers disagree on order, we ranked by severity to the person affected.
Plain English, then the code
Every issue is explained in terms of what a real person experiences. The full report then gives the exact location of every affected element and, where a correct value can actually be computed, that value — a contrast-safe colour verified against your real background, a tabindex correction, an embed title. Where it cannot be computed we say so rather than generating something plausible. Alt text is the clearest example: the same photograph needs different alt text on a banner, a staff directory and a linked card, so a machine cannot know which you meant.
Honest about its own limits
The scan states what it cannot see, and refuses to show a comforting score for a page it could only partially read. That is the opposite of how this category usually markets itself, and it is the only defensible way to sell it.
Pricing
Free to find out whether you have a problem. Paid to get the locations, the recommended action for each and the ability to rescan until it is clean.
See whether you have a problem
Free
Scan any single page and get the risk score, every issue type found, and how many elements each affects — separated into what is proven, what needs checking and what needs a person.
- Any single page, as often as you like
- Risk score with what it does and does not mean
- Every issue type found, with a count
- Confirmed / potential / manual-review classification
- WCAG criterion for each finding
- The full list of what automation cannot check
- No signup, no email gate
- Exact element locations
- Recommended action per element
- Generated values where we can compute them
- Printable report
Professional Accessibility Scanner
$99one-off · 30 days
Identify accessibility issues, understand which WCAG criterion each maps to, and get actionable guidance — without misleading compliance claims.
For: Small businesses, SaaS founders, website owners, freelancers
- One domain, unlimited scans for 30 days
- Every page of that site — scan, fix, rescan to confirm
- Exact location of every affected element
- Recommended action for each, written for a developer
- Computed values where they can be computed — a contrast-safe colour that keeps your brand hue, tabindex corrections, embed titles
- Full WCAG mapping, criterion by criterion
- Manual-review checklist: the named things automation cannot judge
- Printable report (browser print / save as PDF)
- Score tracking across scans
- Saved report history
Accessibility Reporting For Agencies
$149one-off · 30 days
Run scans across multiple client websites and produce professional accessibility reports without paying enterprise software prices.
For: Web, SEO and development agencies, accessibility consultants
- Up to 10 domains, unlimited scans for 30 days
- Everything in Pro, on every one of those domains
- One purchase covers a whole client roster — no per-site top-ups
- White-label reports — your name and logo on the printed report, our brand off it
- Printable per-client report
- Priority support by email
- Multi-domain dashboard
- Team seats
- Historical tracking across scans
- Priority queue processing
Not built yet, and not charged for. If they ship while your 30 days are running you get them at no extra cost; if you need them now, email us and we will tell you honestly where they are.
Why we are not the cheapest
Most tools in this category compete on how many warnings they can generate. A bigger number looks like better coverage, and it is easy to produce: count the same rule once per element, include criteria no software can evaluate, and report every guess as an error.
The problem is what it does to the report. If a list of ninety “errors” contains forty things that were never wrong, you cannot tell which fifty matter — so either you fix all ninety, or you stop trusting the list. Both are expensive. We would rather hand you a shorter list you can act on without second-guessing it, which is why every finding says whether we proved it, suspect it, or need you to go and look.
What you are paying for is the separation. Not a bigger number — a report where the confident parts have earned it.
What each plan includes
| Feature | Free | Pro · $99 | Agency · $149 |
|---|---|---|---|
| Accessibility scan | Yes | Yes | Yes |
| Confirmed / potential / manual classification | Yes | Yes | Yes |
| WCAG criterion per finding | Yes | Yes | Yes |
| Full report with element locations | Limited | Yes | Yes |
| Recommended action per element | No | Yes | Yes |
| Computed fixes where computable | No | Yes | Yes |
| Unlimited scans | No | Yes | Yes |
| Printable / PDF report | No | Yes | Yes |
| Domains covered | 1 page | 1 | 10 |
| Access period | — | 30 days | 30 days |
| Priority email support | No | No | Yes |
| White-label reports | No | No | Yes |
| Multi-domain dashboard | No | No | In development |
| Team access | No | No | In development |
| Historical score tracking | No | In development | In development |
Rows marked “In development” are not built. They are listed so the table cannot be read as implying they exist — a plan you can buy today should not be sold on features that do not.
Scanning more than 10 domains, or want this on a retainer with human audits attached? Email hello@mymindstudio.ai and we will quote against what you actually need.
Learn the thing, not just the score
Plain-English guides to what the law actually requires, what the checks mean, and how to fix each one properly.
- All accessibility guides — the full index
- ADA website compliance, explained without the legalese
- WCAG 2.2 Level AA in plain English
- Why accessibility overlay widgets do not work
- How to write image alt text that actually helps
- Free colour contrast checker — check any two colours against AA and AAA
Frequently asked questions
Is this an ADA compliance certification?
No, and be wary of anything that claims to be one. Automated tools can detect roughly 30 to 40 percent of accessibility requirements. This scan tells you which detectable issues you have and how to fix them, and lists separately what only a human review can assess. It is not legal advice and it is not a certification.
Do accessibility overlay widgets protect me from lawsuits?
The evidence says no. UsableNet, which tracks these filings, counted 1,023 companies running a widget that were sued in 2024 and 1,416 in 2025 — 703 of those in the first half of the year alone. The count covers any accessibility widget, not accessiBe specifically. Installing an overlay is not a recognised legal defence, and major disability organisations including the National Federation of the Blind oppose them because they frequently interfere with the screen readers they claim to help. In January 2025 the FTC ordered accessiBe, the largest overlay vendor, to pay $1,000,000 over claims that its product could make any website WCAG-conformant.
How do I check if my website is ADA compliant?
For most private businesses under ADA Title III there is no detailed website standard to measure against — the Department of Justice has not issued one, so "compliant" is not a state any tool can certify. WCAG 2.1 Level AA is the practical benchmark everyone uses anyway. State and local government websites are different: a 2024 DOJ rule under Title II makes WCAG 2.1 AA their actual legal standard, with deadlines in 2027 and 2028. What you can check today is the shorter list of failures that ADA complaints are actually built on: run an automated scan on your busiest pages, spend ten minutes operating the site with the keyboard alone, then check your video for captions and your forms for labelled fields. That finds most of what a demand letter would cite, and it takes under an hour.
Is an automated accessibility scan enough on its own?
No, and any tool implying otherwise is selling you something. Automated checks reach roughly 30 to 40 percent of what the WCAG guidelines ask for. Whether your alt text is accurate, whether a screen reader can operate your modal, whether focus is visible as you tab — those need a person. A scan is the fastest way to find the provable failures; it is the start of the work, not the end of it.
I received an ADA demand letter about my website. What now?
Talk to a lawyer first — nothing on this page is legal advice, and the letter will have deadlines. In parallel it helps to know what your site actually does wrong, because the specific failures named in these letters are usually a short, fixable list. A scan gives you that list in a few seconds, and it changes the conversation from an unknown accusation to a defined piece of work with a cost attached.
What is an accessibility overlay, and should I install one?
An overlay is a widget you paste into your site that promises to fix accessibility automatically from the outside. The evidence against them is strong: sites running overlays were named in hundreds of ADA lawsuits, the National Federation of the Blind opposes them, and in January 2025 the FTC ordered accessiBe, the largest vendor, to pay $1,000,000 over its claims. Fixing the underlying page is slower and it is the thing that actually works.
Why do you only run 14 checks when other tools advertise hundreds?
Because a check count is a marketing number, not a measure of coverage. A tool reporting "396 checks" is usually counting one rule per element it inspected, or counting every WCAG success criterion including the ones no software can evaluate. Our 14 are the distinct machine-detectable failures that actually appear in filed ADA complaints, ordered by how often they get cited. The full list is published on this page with the WCAG criterion for each, which is something you can verify — a headline number is not.
What does the free preview include?
The risk score, every issue type found with a count of affected elements, and a plain-English explanation of what each one does to a real visitor. That is enough to know whether you have a problem and how big it is. The paid report adds the exact location of every affected element and the corrected value to drop in.
Can I scan more than one page?
Yes. The free scan covers the page you paste. Pro at $99 covers one domain for 30 days with unlimited scans, so you can go through every page of the site and rescan after each fix to confirm it landed. Agency at $149 does the same across up to 10 domains, which is what makes it worth having if you look after client sites.
Why $99 when other scanners are free?
Because the free ones are usually funded by something else — an overlay subscription, an enterprise platform upsell, or your email address. Ours is funded by the report. What you are buying is not a bigger issue count: it is a report where each finding states whether we proved it, suspect it, or need you to check it yourself, plus the exact location and recommended action for every one. If you only need to know whether you have a problem at all, the free scan answers that and we are not going to pretend otherwise.
What is the difference between a confirmed failure and a potential issue?
A confirmed failure is proven by the evidence in your page — an alt attribute is absent, a form field has no label anywhere, a measured contrast ratio is 2.85:1. A potential issue is a real signal where the criterion allows for context we cannot see: generic link text may be rescued by the sentence around it, and a table with no header cells is only a problem if it holds data rather than layout. Manual-review items are a third category we never score at all, because the answer needs a person. They are labelled on every finding and the classification is in the free scan too — telling you that something is unproven is not a feature we would put behind a paywall.
Does the Agency plan include white-label reports?
Yes — this shipped. On the printed report your agency name and logo appear in the header and our branding comes off, so it reads as your document; the one line branding never removes is the disclaimer. Your logo is stored in your own browser and never uploaded to us. Still in development and marked as such: the multi-domain dashboard, team seats, historical tracking and priority queue processing. If those ship while your access is running, you get them at no extra cost.
Which issues actually get cited in lawsuits?
Missing image alt text, unlabelled form fields, controls with no readable name, video without captions, and low-contrast text are the ones complaints are usually built on. Findings here are ordered by that litigation frequency rather than by WCAG chapter, so the first thing on your list is the first thing a plaintiff firm would notice.
Does the ADA actually apply to my website?
Courts have generally held that Title III applies to websites of businesses open to the public, though the reasoning varies by circuit and the DOJ has not issued a general private-sector web rule. In practice it is being decided by volume litigation rather than by settled doctrine: over 5,000 web accessibility suits were filed in the US in 2025 including state courts. If you have received a demand letter, talk to a lawyer rather than to a scanner.
Which WCAG version and level should I be aiming at?
WCAG 2.1 Level AA is the de facto benchmark — it is what most settlement agreements and procurement requirements name, and what Section 508 and EN 301 549 point to. WCAG 2.2 AA adds nine criteria and is the current recommendation. This scanner maps every finding to a specific criterion so you can see exactly which one you are failing.
Will fixing accessibility help my SEO?
Some of it overlaps directly. Image alt text, a declared page language, a real page title, sane heading structure and descriptive link text are all accessibility requirements and all things search crawlers use. The overlap is real but partial — captions and contrast will not move rankings. Treat SEO as a side effect, not the reason.
Why does my site show an incomplete scan?
Because the page assembled itself in the browser rather than arriving as HTML. A single-page app built with React, Vue or a site builder often ships an empty shell plus JavaScript, so a server-side scan sees almost nothing. Rather than showing you a reassuring near-zero score on a page it could not read, the scanner says so. Those pages need testing in a real browser, which we can do as a service.
Can you fix the issues for us?
Yes — that is the actual business. The scanner exists so the conversation starts with evidence instead of a sales pitch. Send us the report at hello@mymindstudio.ai and we will quote the remediation, or book a call and we will walk through it with you.
Nothing on this page is legal advice. Accessibility obligations vary by jurisdiction and by how a site is used. If you have received a demand letter, talk to a lawyer.