Accessibility
Last reviewed 2026-11-07.
Our commitment
Velor targets WCAG 2.2 Level AA on velorplatforms.com, the chat widget and the signed-in product. That is the standard we build to and review against; it is a target rather than an audited conformance claim, and the scope of our checking is stated plainly below rather than implied.
What is automated today. The scheduled end-to-end suite runs axe-core against the sign-in page and asserts zero violations. Separately, and run by hand rather than on a schedule, a crawl of all 82 pages of this site at phone and desktop width runs axe-core with the WCAG 2.0, 2.1 and 2.2 A and AA rules and the last run found no serious or critical violation; a second script (pnpm check:a11y) runs axe in light, dark, reduced-motion and forced-colors modes and also checks 320-pixel reflow, 24-pixel tap targets, landmarks and heading order, and its November 2026 run led to fixes across the site; a keyboard-only script walks six pages, the mobile menu and the chat widget; and component tests run axe over the marketing sections and a sample of dashboard components; and a script renders the signed-in application's pages with sample data, not customer data, runs axe on each at phone and desktop width, and tabs through the first controls of each. None of these is a merge gate.
What is not. Automated rules find only some of what WCAG asks for. Nobody has yet reviewed these pages with a screen reader or with speech input, nobody has run 200% zoom across the site, and the signed-in pages were scanned only with sample data; the ones that need a real record to exist could not be scanned. An earlier version of this statement described broader automated auditing, and a conformance score, that we do not in fact produce; this version replaces it. We would rather state the coverage we have than a figure nobody measures.
Criterion-by-criterion status
We assessed all 55 WCAG 2.2 Level A and AA success criteria against three surfaces. The counts below summarize our criterion-by-criterion table; we send that table, with the test or the stated gap behind each row, to anyone who asks.
| Surface | Supports | Partially | Does not | Not evaluated | Not applicable |
|---|---|---|---|---|---|
| Marketing site and sign-in | 5 | 19 | 1 | 24 | 6 |
| Chat widget | 3 | 15 | 0 | 32 | 5 |
| Signed-in product | 2 | 16 | 1 | 34 | 2 |
“Supports” means a test or a scripted run checked the criterion on that whole surface and found nothing, or the thing it governs does not exist. Most rows are “Partially” or “Not evaluated” because automated rules cover only part of a criterion, or nothing has looked at it yet. The two we know we do not meet: the signed-in product's single-letter shortcuts (2.1.4) can not be turned off, and the home page's looping demo (2.2.2) has no pause control, though your system's reduced-motion setting stops it.
What we've implemented
- Color contrast. Our text colors are chosen against WCAG 2.1 AA (4.5:1 for normal text, 3:1 for large text). The secondary text color is checked by an automated test at 5.0:1 on white, and the chat widget has its own contrast tests. We have not measured every page, so individual pages may still fall short; tell us where.
- Keyboard navigation. Controls are built from native buttons, links and form fields wherever we can. A keyboard-only script confirmed, on the home, pricing, sign-up, sign-in, password-reset and demo-workspace pages, that every control can be reached, focus is always visible, focus is never stuck, the phone menu closes on Escape and returns focus to its button, and the chat widget opens from its launcher and closes on Escape. We have not audited every page by keyboard. Signed-in pages were tabbed through only as far as their first 25 controls, with sample data, and no task has been walked by keyboard end to end.
- Skip link and landmarks. The first Tab stop on every page is a “Skip to main content” link, and the pages checked have a main landmark for it to land on. Code samples and tables that scroll sideways can be focused, so keyboard users can scroll them.
- Semantic HTML. Pages use proper heading hierarchy (H1 → H2 → H3), landmark regions (header, main, nav, footer), and ARIA labels where native semantics are insufficient.
- Screen readers. Menu and drawer controls expose their state. The site's mobile menu button carries
aria-expanded,aria-controls, andaria-label; in the Velor platform, the dropdown menus and the mobile sheet menu are built on a component library that setsaria-expandedandaria-haspopupon their triggers. Decorative images are markedaria-hidden. - Reduced motion. The decorative motion (a drifting background, a looping hero demo, a breathing call button) and the small looping indicators (pulsing text, bouncing dots) all stop when your system asks for reduced motion; a test fails if a new looping animation has no such override. Without that setting the hero demo and the background keep looping and have no pause control, which is a known gap (see below). Spinners stay on because they are the only sign that something is loading.
- Form accessibility. Form fields carry visible labels or
aria-labelattributes, and automated tests check this on the sign-up, knowledge-upload, widget-settings and agent-settings forms. Failed saves are announced to screen readers. Other forms have been fixed as we found problems, not audited as a set.
Known limitations
We do not currently have a formal VPAT (Voluntary Product Accessibility Template, signed off by an independent assessor), and one is not available on request. The table described above is our own self-assessment, not a VPAT. If your procurement process asks for one, contact us: we will answer your accessibility questions in writing, against what is stated on this page.
Known gaps as of this review: the sign-in page and the signed-in product have not been reviewed with a real screen reader. The scripts above are run by hand and are not a merge gate.
The signed-in product has been checked less than this website: its pages were scanned by axe-core with sample data, and the first 25 controls of each were tabbed through, but no signed-in task has been walked by keyboard or with a screen reader.
Chat widget accessibility statement
The widget.js chat widget runs on our customers' websites, in its own shadow DOM, so the host page's CSS and focus management are the customer's and outside our control. We build the panel to the same WCAG 2.2 AA target as the rest of the product.
- Built in. The message list is a live region that is held busy while a reply streams, so a screen reader announces the finished reply rather than every fragment. Controls carry roles and labels. On a phone-width screen the panel is a modal dialog and Tab stays inside it; on a wider screen it is a non-modal dialog. The panel sets its own language for the visitor's locale and lays right-to-left replies out right-to-left. The proactive message button is named with the message itself.
- Checked. Opening the widget from its launcher with the keyboard moves focus into it, Escape closes it and focus returns to the launcher (verified on the built widget by the keyboard script). An axe-core scan of the open panel, with the WCAG 2.0, 2.1 and 2.2 A and AA rules at the default theme, found no violations, including contrast, target size, names and labels. Its own unit tests pin the dialog, live-region, language and button-name behavior above.
- Not checked. The widget is not covered by the axe-core sweep that runs on a schedule; the scan above is run by hand. It has not been tested with a screen reader, at 200% zoom, or with custom text spacing. Colors a customer sets for their own theme are not checked: a low-contrast brand color can fail contrast, and that is the customer's responsibility to review. Treat this as a target with partial evidence, not as an audited conformance claim.
An earlier version of this section asserted conformance for the one component that runs on your site rather than ours; this wording replaces it.
Feedback and contact
If you encounter an accessibility barrier on velorplatforms.com or within the Velor platform, email caden@velorplatforms.com with “Accessibility” in the subject line. We target a 2-business-day response for accessibility reports.