Do accessibility overlays work?
"Get compliant in 48 hours with one line of JavaScript." If you run a website, you have seen the ad. The pitch is an accessibility overlay: a script that claims to detect and repair accessibility failures at runtime, sold as a subscription, marketed as instant protection from lawsuits and regulations. About a billion dollars of these subscriptions are sold on that promise. This guide examines whether the promise holds, using the public record rather than vendor copy or competitor copy, ours included.
What overlays actually do
Technically, an overlay is a third-party script that runs in your visitors' browsers after your page loads. Its features fall into two groups:
- User-facing widgets. A floating button opens a panel where visitors can enlarge text, change contrast, pause animations, or switch to a "screen reader profile."
- Automated remediation. The script inspects the DOM and attempts live repairs: guessing alt text for images, adding ARIA attributes, adjusting tab order, relabeling buttons.
The first group is largely redundant: operating systems, browsers, and assistive technologies already provide zoom, contrast, and motion settings, configured once and applied everywhere. The second group is where the compliance claim lives, and where it fails.
Why runtime repair cannot deliver conformance
Automated detection is genuinely useful; we build on it ourselves. But detection and repair are different problems. Software can detect that an image lacks alt text. It cannot know what the image means, which is what the alt text must convey. It can detect a form field without a label; it cannot know whether the field wants a billing or a shipping address. It can see the DOM order; it cannot know the order a sighted user is intended to read.
Independent evaluations consistently find that automated tooling can identify only about half of WCAG failures, and can correctly repair far fewer. Meanwhile the repairs themselves add risk: injected ARIA that conflicts with the page's real semantics is a new bug, delivered at runtime, differing across browsers and assistive technology combinations. Many screen reader users report disabling or blocking overlays because the "help" interferes with configurations they already rely on.
The deeper problem is architectural. The defective code is still defective; the overlay decorates it. Remove the script, and nothing was ever fixed.
What the public record shows
- Litigation. UsableNet's annual digital accessibility lawsuit reports have documented hundreds of US lawsuits per year filed against companies whose sites were running an overlay at the time of the complaint. Whatever an overlay provides, immunity from legal action is demonstrably not it.
- The FTC. In 2023, the US Federal Trade Commission ordered accessiBe to pay one million dollars for misrepresenting that its AI product could make any website WCAG-conformant, and for failing to disclose paid endorsements. A regulator formally treating the core marketing claim as deceptive is about as direct as evidence gets.
- The practitioner consensus. The Overlay Fact Sheet, an open letter signed by hundreds of accessibility professionals, including many disabled users of assistive technology, asks site owners to adopt a simple position: overlays may offer some usability features, but they do not make a site conformant, and practitioners should not advise clients to rely on them.
- European context. Under the European Accessibility Act, conformance is assessed against EN 301 549, meaning WCAG 2.1 AA in the actual service. French authorities and the European Commission's guidance have both noted that overlay widgets do not constitute conformance with the underlying requirements.
The honest comparison
It is fair to ask how a tool like Complaudax differs, since we also run automated software against pages. The difference is the claim and the architecture. Complaudax injects nothing into your visitors' pages and repairs nothing at runtime. It is a workbench for the people fixing the site: it finds issues (automated detection, powered by axe-core), tracks them per client and site until the code is genuinely changed, structures the manual review that software cannot do, and generates documentation from the verified history.
And the claim we will not make is the one this whole category was built on: no tool makes a site compliant. People fix websites. Tools make the work visible, repeatable, and provable.
If a client is tempted anyway
The overlay pitch wins on apparent economics: a widget subscription looks cheaper than remediation. Three points usually reframe the decision:
- The legal exposure remains, as the lawsuit data shows, because the site itself remains non-conformant.
- Part of the user base the client is trying to serve actively blocks or distrusts the widget.
- Real remediation is a finite project with durable value; a widget is rent, paid forever, for the appearance of progress.
A pragmatic counter-offer: a baseline audit, a prioritized fix plan, and a documented re-audit. On Complaudax's free plan the baseline audit costs nothing to produce, which makes the honest path easy to quote and easy to compare.
So, do overlays work? As a set of optional usability preferences, mildly, and redundantly. As a route to conformance, the evidence from users, courts, regulators, and the practitioner community all points the same way: no.