The latest articles on DEV Community by OverlayRiskWitness (@overlayrisk). https://dev.to/overlayrisk https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4008931%2F658aa13c-34c4-4a09-a726-b1e772e9026f.png https://dev.to/overlayrisk en OverlayRiskWitness Tue, 30 Jun 2026 08:59:10 +0000 https://dev.to/overlayrisk/how-to-test-what-an-accessibility-overlay-actually-changes-12mm https://dev.to/overlayrisk/how-to-test-what-an-accessibility-overlay-actually-changes-12mm <h1> How to test what an accessibility overlay actually changes </h1> <p>Accessibility overlays are easy to install and hard to evaluate independently.</p> <p>A site can add an overlay widget, display an accessibility statement, and still have no neutral evidence of what the overlay changed on the page. The vendor dashboard is useful, but it is not independent of the vendor.</p> <p>The cleaner test is a before/after experiment.</p> <h2> The experiment </h2> <p>Take one public page.</p> <p>Load it once with the overlay active.</p> <p>Load it again with the overlay blocked.</p> <p>Run the same automated checks on both versions.</p> <p>Then compare the result.</p> <p>That gives you a narrow but useful question:</p> <p>What changed when the overlay was present?</p> <h2> What the diff can show </h2> <p>The result should not say "compliant" or "not compliant."</p> <p>That is too broad for an automated check.</p> <p>Instead, the result should say things like:</p> <ul> <li>this axe-core rule failed with the overlay blocked and still failed with the overlay active</li> <li>this rule changed state when the overlay loaded</li> <li>this page could not be tested reliably</li> </ul> <p>The useful artifact is the evidence trail: page URL, overlay state, timestamp, screenshot, DOM snapshot, and finding.</p> <h2> Why public claims matter </h2> <p>Many sites running overlays also publish accessibility statements. Some point to WCAG. Some imply ADA risk is handled. Some rely on the overlay vendor's own language.</p> <p>That public claim is the context. A before/after witness can quote the claim and show what the page-level automated evidence observed.</p> <p>It still does not decide legal compliance.</p> <p>It gives the site owner and counsel something concrete to inspect.</p> <h2> Limitations </h2> <p>axe-core does not catch every WCAG issue. Automated checks are not a substitute for manual expert audits or user testing.</p> <p>This method is not a legal opinion.</p> <p>It is a way to stop arguing from marketing claims and inspect the page itself.</p> <h2> Try it </h2> <p>I built OverlayRiskWitness to run this experiment on one page for free:</p> <p><a href="https://overlayrisk.com/?utm_source=devto&amp;utm_medium=article&amp;utm_campaign=launch-2026-06-30&amp;utm_content=technical_walkthrough#run-witness" rel="noopener noreferrer">https://overlayrisk.com/?utm_source=devto&amp;utm_medium=article&amp;utm_campaign=launch-2026-06-30&amp;utm_content=technical_walkthrough#run-witness</a></p> <p>Paste a URL from a site running accessiBe or UserWay. The product loads the page with the overlay on and blocked, runs the comparison, and shows one timestamped finding before signup or payment.</p> <p>Evidence documentation, not legal advice.</p> a11y webdev testing playwright