Compliance and legal teams
In companies selling to consumers in the EU, who have a legal obligation, an exposure, and a board asking about it.
Accessibility
A written WCAG 2.2 audit combining static analysis, a real Chromium browser and axe-core. Remediation a person approves. VPAT and statement generation in seven languages. Monitoring that catches regressions.
If you are here to find out what the law requires rather than what we sell, that is a separate page and it does not try to sell you anything.
This is the most important sentence on the page, and it works against our own product category, so it is worth being direct about it.
A widget that sits on top of an inaccessible website does not make the underlying content accessible. It cannot invent a meaningful alternative for an image that has none, and it cannot give structure to a page that has none. Overlays have been publicly criticised by courts and by organisations representing blind and partially sighted users, and that criticism is largely deserved.
A widget is a convenience for the visitor — larger text, adjusted contrast, reduced motion, a language they read. It is not a substitute for building the interface properly. We ship one because visitors use it. We do not sell it as conformity.
The audit runs three complementary methods across your main pages: static analysis of the HTML, a real Chromium browser driven by Playwright, and axe-core. On top of those sit 80 checks written specifically to cover what axe-core does not — 35 on the static HTML and 45 that need the live browser, including walking the entire page with the keyboard to find focus traps. The result is a written report with a score calculated by a published formula you can reproduce.
This is the stage the platform does not do for you, and it is worth being blunt about it. Contrast, focus, labels, structure, keyboard traps and alternative text are changed in the interface itself — by your own developers, or by us under a separate engagement quoted from the audit. What the report hands over is exactly what to change and where: the element, the page, the measured value and a suggested replacement. Claude, Gemini or GPT draft the parts that are language work — alternative text, ARIA labels, simplified wording — and a person reviews those drafts before they go into the report. Nothing is ever written to your site. A scanner that edits pages by itself is how a site ends up with hundreds of confident, wrong image descriptions.
Conformity you cannot document is conformity you cannot defend. The platform generates, in seven languages: an accessibility statement, a VPAT / Accessibility Conformance Report, the scan report, and a conformance badge. Documents come out in the client’s own language, not only in English — which matters when the authority reading them is national.
Accessibility regresses with every release. A state proven once says nothing about the same site six months later. Scheduled scans, comparison between runs, email reports and webhooks turn a one-off document into a measurement you can trend.
Keeping it
Accessibility regresses with every release. The panel keeps every scan of every domain, so the question stops being “are we compliant?” and becomes “which of our sites moved, and in which direction?”
Still moving
Including the entry where we ran the scanner on ourselves and did not like the result.
Recently shipped in accessibility
Automated testing detects a portion of the WCAG criteria reliably — missing alternatives, contrast failures, structural problems, keyboard traps. Criteria that depend on meaning and context need a person. Whether an alternative text is useful is not a question code can answer.
Our report says so on its first page: it is an automated preliminary check, and it does not replace a manual evaluation. In a market where “compliance scores” are handed out freely, that sentence is the reason to trust the rest of the document.
In companies selling to consumers in the EU, who have a legal obligation, an exposure, and a board asking about it.
Subject to accessibility requirements for longer, and usually needing the statement and the conformance report more urgently than they need the code fixed.
Managing sites for their own clients. There is a partner panel with cascading branding, so the platform runs under your name.
The accessibility platform has its own site — accessibility.valenys.com — published in six languages: English, Romanian, German, French, Spanish and Italian. It includes a free scan of one page: enter an address and you get a WCAG score, the critical issues and what to change first, with no account and no card.
That scan is an automated preliminary check on one page, not the audit described above. It is, however, the cheapest way to find out whether you are looking at a small problem or a large one.
It converts an unknown into a number, a list and a plan, in days rather than months — and it is the only sensible input to a remediation budget.
No, and neither can anyone else honestly. Conformity depends on content that changes daily, on third-party components, and on criteria that require human judgment. What we can do is measure against WCAG 2.2, tell you precisely what to change and where, document the state accurately including what is not fixed, and keep measuring. The changes themselves are made in your code — by your developers, or by us as a separate piece of work.
Days rather than months, depending on how many pages and journeys are in scope. The scope is agreed before we start, because “the whole site” and “the twelve pages that carry your revenue” are very different pieces of work.
The automated audit uses a real browser and a keyboard walk, which catches a class of problems that static analysis misses entirely. Testing with screen reader users is a separate manual engagement — and where the criteria that need human judgment actually get answered.
It is a structured document describing how a product measures against accessibility standards, in a format recognised internationally. Procurement teams and public sector buyers increasingly ask for one. The platform generates it in seven languages.
Not for measurement, and rarely for the statement and monitoring. It can limit what is fixable — some of it lives in a theme or a third-party component you do not control, and no scanner and no widget can reach it either. The audit says which findings those are, so you know what is a decision and what is a constraint.
Yes. The partner panel supports agencies managing their own clients, with branding cascading down to the widget.