1. What this statement says, and what it refuses to say
This statement is split three ways on purpose: what is measured and by which gate, what is known to be wrong, and what has not been measured at all. Anything in the third group is unknown. It is not "probably fine", and this page will not describe it that way.
Every measured line below names the file or the command that measures it. This software is open source: you can read the gate and run it yourself rather than take the line on trust.
A conformance claim against an accessibility standard We do not have this
None is made here. A sentence such as "WCAG 2.1 AA compliant" means that the whole product has been audited against that standard by someone qualified to do it, and recorded. That has not happened for this service. Writing it anyway is the most common untruth on pages of this kind, and it is the one thing this page exists to avoid.
2. What is measured, and by which gate
These run against a real layout engine or a real browser, not a simulated one. That distinction matters: the repository's unit tests run in an environment that computes no layout at all, where no box has a height, nothing overflows and no touch target can be measured. Questions about size and overflow cannot be asked there and this statement does not pretend they were.
Touch target size, horizontal overflow and clipping on a narrow screen Measured
Pages are opened in a real browser at 320 by 480 and again at wider sizes, and checked for: horizontal overflow, any interactive element clipped past the edge, touch targets smaller than 44 by 44, neighbouring targets closer than 8 pixels apart, silently truncated text, and content squeezed into a narrow strip by nested padding. The same run is repeated with the layout reversed, for the two supported languages that read right to left. Every interface component is measured this way on every change; the assembled corporate pages are measured with the same tool when they change, which is not the same thing and is why it is said here rather than rounded up.
Measured by
scripts/mobile-ux-audit
Reduced motion — the rule Measured
When the device asks for reduced motion, the page has to be complete rather than diminished: the same sections, the same links, the same headings, and nothing hidden behind an entrance animation. A gate reads the stylesheets and fails if any declaration that starts motion sits outside the reduced-motion guard, so a single forgotten line cannot slip through by looking harmless.
Measured by
tests/Feature/PublicSite/SceneContractTest.php
Reduced motion, increased contrast and forced colours — visual signatures Measured
The visual gate captures selected pages and display states in a browser, with screenshots 900 milliseconds apart. It compares coarse 16×12 brightness signatures: static checks fail when signature movement exceeds zero, while reference comparisons allow a 12/255 difference. These checks detect some motion and visual regressions; they do not compare every pixel or prove that every accessibility requirement is met.
Measured by
scripts/scene-visual-gate
Text contrast in the corporate palette Measured
Configured palette pairs are computed from the stylesheet and checked against their specified minimum ratios, including 7:1 for the configured body-text pairs. This does not cover every rendered combination or every local desktop colour token. Semi-transparent colours depend on the background behind them and are not covered by these opaque-token checks.
Measured by
tests/Feature/PublicSite/CorporateIdentityContrastTest.php
Right-to-left layout Measured
Two of the nine languages the software supports are written right to left, and every physical direction in a stylesheet is reversed for them: a left padding lands at the end of the sentence rather than the start. A gate scans the stylesheets, the templates and the help articles for physical direction properties and classes, and requires the logical equivalents instead.
Measured by
scripts/logical-direction-gate
Roles, names and keyboard behaviour of the interface components Measured
Every story of every catalogue component is rendered and run through an accessibility rule engine, and the build fails on any violation. Colour contrast is switched off in that run and measured separately at token level, because the environment it runs in computes no colour.
Measured by
resources/js/design-system/a11y.guard.test.tsx
Minimum touch target in the component catalogue Measured
A separate contract holds the interface components to the 44 by 44 minimum, so a control does not fall below it between the browser measurements.
Measured by
resources/js/design-system/touch-target.contract.test.tsx
3. What is known to be wrong
These are measured findings that are still open. They are listed with the same weight as the measured lines above, because a gap that is only recorded in an internal file is a gap the reader never learns about.
Links inside a sentence are smaller than 44 by 44 Known gap
Links written into the middle of a sentence are the height of the line of text they sit in rather than 44 pixels: "Contact us" in the shared block that shows the plans, which appears on every page carrying that block, and "Write to us" on the help page. Making them 44 pixels tall would break the paragraph they are part of. The finding is left open and named here rather than closed by a rule that hides it — and it is written as a shape rather than a list of pages, because a list would be out of date the next time that block appears somewhere new.
Measured by
scripts/mobile-ux-audit
Two links sit closer together than the minimum gap Known gap
On the pricing page, two neighbouring links are 7 pixels apart on a narrow screen where the minimum is 8. Seven pixels rather than eight is a small number, and it is written here for the same reason the larger gaps are: a threshold that is quietly rounded down once is a threshold that stops being one.
Measured by
scripts/mobile-ux-audit
Small targets frozen in the component catalogue Known gap
A checkbox control and a text link in the component catalogue are recorded as below the minimum, together with some layouts that waste horizontal space on a narrow screen. They are held in a baseline file so the count can only go down, never up — but they are debt, and the baseline is a record of debt rather than a pass.
Measured by
scripts/mobile-ux-audit.baseline.json
Limits of automated coverage Known gap
The CI workflow runs the scene visual gate for selected pages and states, alongside automated checks, the logical-direction scan and component catalogue checks. That coverage is bounded: it does not exercise every page, every browser or every user journey, and a passing run does not prove that all visual or accessibility defects are absent.
Measured by
.github/workflows/ci.yml
4. What has not been measured
Nothing in this section should be read as working. It should be read as unknown.
Use with a screen reader by a person who uses one Not measured
No session with a screen reader user has been run against this service. Automated rule engines check the parts of a screen reader experience that a machine can see — that a control has a name, a role and a reachable label. They do not check whether the resulting page is usable, and nothing here has checked that.
iOS, Safari and VoiceOver Not measured
Every browser measurement behind this page was taken in one browser engine, on one machine, and there is no Apple device in that setup. Several things this interface relies on behave differently there. Nobody has looked, so the answer is unknown rather than likely.
Speech input, switch access and screen magnification Not measured
None of these has been exercised against this service.
The rest of the corporate site Not measured
The browser measurements cover the pages that are published today. The site map holds several hundred further addresses that are written but not published; none of them has been measured, and none of them is reachable, so this statement covers what you can actually open and says nothing about the rest.
The workspace application and the guest menu Not measured
The component catalogue behind those screens is measured, component by component. The assembled screens are not measured as whole pages in a real browser the way the corporate pages are. A page can fail where none of its parts does.
5. Tell us what is broken
A report that names the page, the device and what happened is worth more than any gate on this list, because the gates only measure what somebody thought to measure.
How to report something Measured
Use the contact form at /contact. It needs no account, because somebody who cannot use the interface may not be able to sign in to report that. It stores the message and gives you a reference to quote. If the form itself is the thing that is broken for you, the seller's e-mail address is on the About page at /about.
Measured by
app/Http/Controllers/PublicSite/StoreContactMessageController.php
How quickly you are answered We do not have this
No support response time is committed by this service, so this statement does not promise one. A phrase such as "as soon as possible" would be a promise nobody made, and it is left out for the same reason the availability figure is.
The one timing this service does commit to Not measured
The incident notification time has not been entered for this installation, so there is no committed timing to point at here either.