Accessibility Statement

What we have measured on the public AiLabrix pages, how we measured it, and what is still open. Written to be checkable rather than reassuring.

Last updated: 2026-07-26 · v1.0

✓ Conformance target: WCAG 2.1 level AA The target applies to the public marketing pages: the home page in all five languages, plus /pricing, /trust, /compliance and /contact. The authenticated application is not covered by this statement.

→ Current status: partially conformant Every success criterion we can verify by automated measurement currently passes. One verification step is outstanding — testing with a real screen reader — so we describe the pages as conformant on the automatically verified criteria rather than as fully conformant. We would rather state the narrower claim than the one we cannot yet evidence.

What was measured, and how

Two harnesses in the repository render the pages in Chromium and measure the result, so any claim below can be re-checked and any regression is caught rather than discovered by a user:

tools/audit_landing.py — contrast of every text element against its effective background, heading outline, accessible names, duplicate ids, running animations, and video behaviour. Run in English, Italian, and with prefers-reduced-motion: reduce.
tools/audit_wcag_reflow.py — reflow at a 320 px viewport, text resized to 200%, the WCAG text-spacing overrides, keyboard tab order, form labels, and focus-indicator contrast, across all five pages.

Both compare against a pinned baseline and fail if a number gets worse. The statically decidable half is additionally asserted in tests/test_marketing_page_contract.py, which runs with the normal test suite.

Criteria verified

1.4.3 Contrast (minimum) — no text below the required ratio. Two colours were changed to reach this: a division tag measured 3.77:1 and a badge that sat on a semi-transparent background over a video, where the ratio depended on the frame behind it.
1.4.10 Reflow — no horizontal scrolling at 320 px, and no content pushed out of reach. Three separate causes were fixed: a footer grid that never collapsed to one column, cards holding a chart canvas that refused to shrink, and a responsive grid that had been frozen into a fixed one.
1.4.4 Resize text — usable at 200% text size.
1.4.11 Non-text contrast — the focus indicator. It was measured at 1.12:1 against the dark footer, meaning a keyboard user could not see where they were; it is now a two-tone ring that contrasts on either side.
1.4.12 Text spacing — no loss of content with the WCAG line, letter, word and paragraph spacing applied.
1.4.13 Content on hover or focus — navigation menus are dismissible with Escape and stay open while the pointer is over them.
1.3.1 Info and relationships — one h1 per page and no skipped heading levels. Four decorative headings inside an illustrative interface mock-up were turned into non-heading elements.
2.1.1 / 2.1.2 Keyboard — every control is reachable and focus never becomes trapped. Three decorative buttons inside the mock-up were removed from the tab order.
2.2.2 Pause, stop, hide — the looping demo video has a play/pause control, and it does not start on its own when the visitor has asked for reduced motion. Decorative animations stop under the same preference.
1.2.1 Audio-only and video-only — the silent demo video has a text description next to it.
3.3.2 Labels or instructions — every form field on the contact page has a programmatic label.
1.1.1 Non-text content — every image has an alternative, and the interface mock-up is exposed as a single labelled image rather than as dozens of meaningless fragments.

Known limitations

⚠ Not yet verified with assistive technology The measurements above are automated. They cannot tell us whether the reading order and announcements actually make sense to someone using a screen reader. Until that test is done and recorded here, treat the automated result as necessary but not sufficient.

Illustrative interface mock-up. The animated browser window on the home page is presented as a single image with a text description. Text inside it is decorative and is not intended to be read; it is not an accessible substitute for the product itself.
Wide data tables. The sub-processor table and the standards table scroll horizontally inside their own container. This is permitted for tabular data, but it does require horizontal scrolling of that region on a narrow screen.
Third-party accessibility widget. The pages load an open-source user-side adaptation widget (high contrast, font size, animation pause and more) from a pinned version with an integrity check. It is an addition to the work above, never a substitute for it.
WCAG 2.2. Our target is 2.1 AA. Some 2.2 criteria — notably target size (2.5.8) — are not yet met everywhere: a few footer links are below 24 px in height.
Level AAA is not a target. Body text at the secondary grey passes AA and does not pass AAA.
The authenticated application (dashboard, review gates, administration) has not been audited to this standard yet.

Reporting a barrier

If something on these pages blocks you, please write to [email protected] with the page address and what happened. We will reply within five working days. If you use assistive technology, telling us which software and version helps us reproduce the problem.

How this statement is kept honest

The claims above are generated from measurements, not from intent. If a change regresses one of them, the harnesses fail against their baseline and the statement is wrong until the code is fixed — not until someone remembers to edit this page. Any criterion we have not measured is listed under limitations rather than left unmentioned.