How it works
The before/after method, in five steps
OverlayRiskWitness does one thing precisely: it isolates what an accessibility overlay actually changes on your page, and turns that into a timestamped, downloadable record. No score to argue with — just two axe-core runs and the difference between them.
Run the page twice — overlay on, then off
You submit a URL. A hosted headless browser loads the page with the detected accessibility overlay active and runs axe-core, the open-source accessibility rule engine. It then loads the same page with the overlay blocked and runs axe-core again. Two renders, same page, one variable changed: the overlay.
Diff the results rule by rule
We compare the two axe-core runs and record the difference for each rule — what the overlay appears to fix, what it does not touch, and anything it makes worse. This rule-by-rule delta is the core of the witness: it isolates the overlay's real effect from the page's underlying markup.
Capture timestamped exhibits
For each finding we capture a DOM snapshot and a screenshot, each stamped with the time of the run and tied to the specific page URL. These exhibits are the reproducible record — the thing that answers "what did this page actually do, and when" without anyone reconstructing a browser session from memory.
Package the Risk Packet
The findings, the rule-by-rule diff, and the exhibits are assembled into a downloadable Risk Packet. For paid packets we also read your public accessibility statement and quote the relevant compliance language back against what the page does, so the claim and the evidence sit side by side.
Watch for drift over time
Sites change. A redesign, a CMS update, or an overlay configuration change can quietly alter the before/after picture. Drift Monitor re-runs the witness on your chosen pages on a recurring schedule and alerts you when the result moves, so a new gap does not stay invisible until a demand letter finds it.
What a Risk Packet contains
The packet is built to be reproducible and self-explanatory, so counsel or a colleague can read it without rerunning anything.
- Page record
- The exact URL tested and the run timestamp, so every finding is anchored to a specific page at a specific moment.
- axe-core results (on / off)
- The full rule output from both renders, so the comparison is auditable rather than a single summarized score.
- Rule-by-rule diff
- The per-rule delta between overlay-on and overlay-off, isolating what the overlay actually changed.
- DOM snapshot + screenshot
- Visual and structural evidence for each finding, captured at run time and stamped.
- Claim quotation
- The relevant lines of your public accessibility statement quoted next to what the page does (paid packets).
Why on/off instead of a single scan
A normal accessibility scan tells you the page has violations. It cannot tell you whether the overlay you are paying for is responsible for fixing them — or whether the overlay is hiding problems it never actually solved. Running the page with the overlay off is the only way to attribute cause. The diff is the evidence; the timestamp and snapshot make it hold up later.
One page, no signup. See one real before/after finding free. Full packet is $49. This is evidence documentation, not legal advice.