Home › Accessibility overlays
Is accessiBe enough? What an accessibility overlay can and cannot do
The evidence on overlays: what the script changes, what the Overlay Fact Sheet documents, and the lawsuit counts from 2022 to 2026.
Short answer: no. accessiBe, UserWay, AudioEye's widget and every other accessibility overlay run as a script on top of your pages. They can change what a visitor's browser shows while the script runs; they cannot make the site's code meet WCAG, and they have not kept anyone out of court. In January 2025 the FTC ordered accessiBe to pay $1 million over its compliance claims, and roughly one in four sites sued for web accessibility in 2025 and early 2026 had an overlay installed.
What an accessibility overlay is
The Overlay Fact Sheet, a community document signed by more than 1,000 accessibility practitioners, defines overlays as third party code, typically JavaScript, added to a website to improve its accessibility. Every product in the category has two layers:
- The widget. A floating button that opens a panel of adjustments: text size, contrast, spacing, cursor, animations, a reading mask, a "screen reader" or "keyboard navigation" profile.
- Automated remediation. Code that scans the page after it renders and rewrites parts of the DOM: adds ARIA roles and labels it infers from context, generates alt text with an image recognition model, adjusts tab order and tries to manage focus in menus and pop ups. This is the layer that the "compliance in 48 hours" marketing rests on.
Our accessibility widget explainer covers the first layer in detail, including what the free tiers give you. This page is about the second layer and the evidence.
What an overlay actually changes
The remediation runs in the visitor's browser, on each page load, after your HTML has arrived. Three consequences:
- The change is temporary and conditional. Block the script (ad blockers commonly do), let it fail to load, or view the page source, and none of the repairs exist. Your code is exactly as inaccessible as before.
- The repairs are guesses. The script has no idea what your image is for, what a custom dropdown is supposed to announce, or which of three adjacent text nodes is the field's label. It applies heuristics. On a simple brochure site the heuristics are often right; on a product listing page with filters, a cart drawer and a third party checkout they are frequently wrong, and a wrong ARIA label is a new barrier, not a fix.
- The visitor's own tools get in the way. A blind visitor arrives with JAWS, NVDA or VoiceOver already running. The overlay's "screen reader profile" adds a second layer of announcements and keyboard handling on top of theirs. This is the complaint users repeat in the "in their own words" section of the fact sheet.
What overlays cannot do, per the Overlay Fact Sheet
The fact sheet's technical section lists the categories where automated repair does not work: the accuracy of alternative text, form labels, keyboard access and focus control, content in PDFs, canvas, SVG and media, and pages driven by JavaScript frameworks whose content changes after the script has run. It also documents a privacy issue: overlays that remember a visitor's chosen disability profile in a cookie or local storage are processing health related data, which under Article 9 of the GDPR requires an explicit legal basis. Its signatories, who include editors of the WCAG, ARIA and HTML specifications and accessibility staff at large technology companies, commit to four things: never to advocate overlays that are marketed as automated compliance, to advocate fixing errors at the source, to speak up against deceptive marketing, and to encourage the removal of overlays in favour of permanent strategies.
The evidence, 2021 to 2026
| When | What happened | Source |
|---|---|---|
| June 2021 | The National Federation of the Blind's board found that accessiBe "engages in behavior that is harmful to the advancement of blind people" and revoked its convention sponsorship; the convention adopted Resolution 2021-04 against overlays on 10 July 2021. | NFB |
| 2022 | 575 US web accessibility lawsuits were filed against sites with a widget installed. | UsableNet 2023 report |
| 2023 | 933 such lawsuits, a 62 percent increase, out of 4,605 total. UsableNet: many complaints list the widget features themselves as barriers. | UsableNet 2023 report |
| 2024 | 1,023 companies with a live widget were sued, out of more than 4,000 cases. | UsableNet 2024 report |
| 3 January 2025 | The FTC announced an order requiring accessiBe Inc. and accessiBe Ltd. to pay $1 million; the agency alleged the company falsely claimed accessWidget could make any website WCAG compliant and presented paid reviews as independent. The order bars unsubstantiated claims that an automated product can achieve or maintain WCAG compliance. | FTC |
| January to June 2025 | Lawsuits naming widget sites every month: 85, 119, 132, 105, 119 and 99. | UsableNet midyear 2025 |
| 2025 full year | 28 percent of all digital accessibility lawsuits targeted sites using overlays, up from 25 percent in 2024, in a year with more than 5,000 filings. | UsableNet, via Accessibility.Works |
| Q1 2026 | 265 of 1,037 lawsuits (25.55 percent) targeted sites with a widget, up from 209 in Q1 2025. | EcomBack |
Nothing in that timeline requires interpretation. The share of defendants with an overlay has sat at a quarter or more for three years while filings rose. If overlays prevented lawsuits, that number would fall.
Why sites with an overlay still get sued
The people who file these cases are professional testers with a screen reader and a scanner, not casual visitors who might click the accessibility icon. The full year by year filing counts are in the ADA lawsuit guide; here is what they find on an overlay site:
- The scanner still reports the failures, because it tests the rendered code and the overlay's repairs either did not run, ran late, or were wrong.
- The screen reader still cannot complete the checkout, because the checkout is a third party iframe the overlay cannot reach.
- Alt text generated by an image model describes the wrong thing, which is cited as a barrier in its own right.
- The overlay's own panel is not keyboard operable, or covers the "Place order" button at mobile widths, which gives the complaint one more paragraph.
There is a further problem for the defence: the overlay is evidence that you knew accessibility was a legal issue and chose a script instead of fixing the code. It is not a good look in a settlement negotiation.
Is accessiBe enough for anything?
Being precise rather than absolute: an overlay's preference panel can be a convenience for some users, and its automated layer may mask a handful of trivial defects on a simple site. What it cannot be is your compliance strategy, your response to a demand letter, or the basis for the word "compliant" in your accessibility statement. If you keep one, keep it honest: no compliance claims on your site, the panel dismissable and never over content or controls at 375 pixel widths, tested with NVDA to confirm it does not break the checkout, and the budget for the real work protected. If you are replacing it, our accessiBe alternatives comparison shows the options side by side.
What "enough" actually looks like
- Find the real failures. The free scan runs the same class of engine the plaintiffs' scanners use and lists the failing elements with their code. Do the home page, a product page and the checkout.
- Fix the six categories that are 96 percent of detected errors. WebAIM's 2026 study: low contrast (83.9 percent of home pages), missing alt text (53.1), missing form labels (51), empty links (46.3), empty buttons (30.6), missing language (13.5). Test pairs in the contrast checker; list every image with the alt text checker.
- Do the keyboard walk and one screen reader pass on the checkout. Automated tools found 41 percent of planted barriers in the UK government's test; the rest is this step.
- Publish a truthful accessibility statement with a monitored contact and the real known limitations.
- Monitor. Scheduled re-scans, because the next theme update will undo something.
Overlay versus remediation versus AdaClear
| Capability | Overlay (accessiBe, UserWay, AudioEye widget) | Manual remediation | AdaClear |
|---|---|---|---|
| Changes your code | No, runtime only | Yes | Finds the failures; you or Expert Fix change the code |
| Injects a script into your pages | Yes, on every page | No | No |
| Alt text | Machine guessed | Written by a person | Every image listed with its alt; Expert Fix writes them |
| Keyboard and checkout iframes | Cannot reach | Fixed at source | Reported; fixed at source |
| Evidence for a demand letter reply | Complaints cite the widget | Remediation log | Dated scan reports before and after |
| Regression protection | Re guesses each load | Only if you re-test | Scheduled re-scans |
See what the overlay did not fix
Free 30 second scan of any page: score, the failing elements and their code. Nothing injected.
Scan my site free →Frequently asked questions
Is accessiBe enough to make my website ADA compliant?
No. accessiBe and other overlays run a script in the visitor's browser that adjusts the page after it loads. The HTML your site serves is unchanged, so a scanner, a screen reader user and a plaintiff's tester see the same failures. In January 2025 the FTC ordered accessiBe to pay $1 million over claims that its widget made sites WCAG compliant, and about 28 percent of accessibility lawsuits in 2025 named sites with an overlay.
What is an accessibility overlay?
Third party code, usually a single JavaScript tag, added to a website to improve accessibility. It has two parts: a visible widget with adjustments such as text size and contrast, and an automated remediation layer that guesses ARIA labels, alt text and focus behaviour at runtime. The Overlay Fact Sheet, signed by more than 1,000 practitioners, documents what that automation can and cannot do.
Why do accessibility experts oppose overlays?
Because the automated repairs are guesses that are often wrong, because the widget interferes with the screen readers and settings that disabled users already run, because some overlays store a visitor's disability profile in ways that raise GDPR issues, and because the products are marketed as compliance. The National Federation of the Blind adopted a resolution against overlays in July 2021 and revoked accessiBe's convention sponsorship.
Do sites with an overlay still get sued?
Yes, at a steady rate. UsableNet counted 575 such lawsuits in 2022, 933 in 2023 and 1,023 in 2024; overlays were on 28 percent of sites sued in 2025 and, per EcomBack, on 25.55 percent (265 of 1,037) in the first quarter of 2026. Many complaints cite the widget's own features as barriers.
Can I keep an overlay as a convenience feature?
You can, if you make no compliance claim based on it, the panel can be dismissed and never covers content or the checkout button on small screens, you have tested that it does not break the site with NVDA or VoiceOver, and your code independently passes WCAG. Under those conditions it is a preference panel, not a strategy.
What should I do instead of an overlay?
Scan the pages a tester would use, fix the six failure types that make up 96 percent of detected errors (contrast, alt text, form labels, empty links, empty buttons, page language), test the checkout with a keyboard and a screen reader, publish an honest accessibility statement with a contact, and re-scan after every release. AdaClear does the scanning and monitoring without injecting anything into your site.
Sources
- Overlay Fact Sheet (community statement, more than 1,000 signatories): https://overlayfactsheet.com/en/
- FTC press release, 3 January 2025: order requires accessiBe to pay $1 million over deceptive WCAG compliance claims: https://www.ftc.gov/news-events/news/press-releases/2025/01/ftc-order-requires-online-marketer-pay-1-million-deceptive-claims-its-ai-product-could-make-websites
- National Federation of the Blind, National Convention Sponsorship Statement Regarding accessiBe (June 2021): https://nfb.org/about-us/press-room/national-convention-sponsorship-statement-regarding-accessibe
- National Federation of the Blind, 2021 Resolutions (Resolution 2021-04 on overlays): https://nfb.org/resources/speeches-and-reports/resolutions/2021-resolutions
- UsableNet, 2023 Year-End Report on ADA Digital Accessibility Lawsuits (933 suits against widget sites, 4,605 total): https://3280432.fs1.hubspotusercontent-na1.net/hubfs/3280432/2024-ADA-Report-Digital-Accessibility-Lawsuits.pdf
- UsableNet, 2024 Year-End Report (1,023 suits against widget sites, 4,000+ total): https://3280432.fs1.hubspotusercontent-na1.net/hubfs/3280432/UsableNet-2024-Year-End-ADA-Web-Lawsuit-Report.pdf
- UsableNet, 2025 Midyear Accessibility Lawsuit Report (2,019 suits in H1 2025, monthly widget counts): https://blog.usablenet.com/2025-midyear-accessibility-lawsuit-report-key-legal-trends
- Accessibility.Works, Accessibility Overlay Widgets Attract Lawsuits (28 percent in 2025, up from 25 percent in 2024, citing UsableNet): https://www.accessibility.works/blog/accessibility-overlay-widgets-attract-lawsuits/
- UsableNet, ADA Web Lawsuit Trends for 2026: What 2025 Filings Reveal (5,000+ suits, 1,427 repeat defendants): https://blog.usablenet.com/ada-web-lawsuit-trends-2026
- EcomBack, Q1 2026 ADA Website Lawsuit Report (1,037 suits; 265 against sites with widgets): https://www.ecomback.com/ada-website-lawsuits-recap-report/q1-2026
- WebAIM Million 2026: accessibility of the top 1,000,000 home pages: https://webaim.org/projects/million/
- GOV.UK accessibility blog, What we found when we tested tools on the world's least-accessible webpage (best tool found 41 percent of 143 barriers): https://accessibility.blog.gov.uk/2017/02/24/what-we-found-when-we-tested-tools-on-the-worlds-least-accessible-webpage/
- Regulation (EU) 2016/679 (GDPR), Article 9 on special categories of personal data: https://eur-lex.europa.eu/eli/reg/2016/679/oj
- W3C, Web Content Accessibility Guidelines (WCAG) 2.2: https://www.w3.org/TR/WCAG22/