Accessibility

Accessibility statement

Standing Proof is built against WCAG 2.1 AA as its design target. This page says what that means concretely: what has actually been measured, what is practiced but not yet independently audited, and what we know is not right yet.

What has been measured

Every route the audit lists (the list is in the repository and grows with the product) can be walked by an instrumented script at phone widths (320px and 375px) against two hard rules: no page may scroll horizontally, and every interactive control must meet WCAG 2.5.8’s minimum target size of 24×24 CSS pixels. It runs every night the product has changed, against the demo, at both widths, and a night on which either rule is broken is a failed run. It is not part of the check a change must pass before it is merged, so a regression can reach the product and be found the next night. A run on 2026-10-03, the first in a month, found no page that scrolls sideways, and found text links, buttons and disclosure rows shorter than 24 pixels on pages across the product. They were fixed on 2026-10-04, and the audit then ran clean at both widths on every route it lists.

What is practiced throughout

Semantic HTML with a heading hierarchy; a skip-to-content link on application screens; visible labels on form fields rather than placeholder-only inputs; text alternatives on images written to describe what is in the frame; icon-only controls carrying accessible names; and horizontally scrolling regions (wide tables) that are keyboard-focusable with a visible cue that more content exists. The 320px measurement above doubles as reflow evidence: content that lays out without horizontal scrolling at 320 CSS pixels is what WCAG’s reflow criterion asks of a desktop page at 400% zoom. Status indicators pair color with a text label rather than relying on color alone.

Known limitations

Many controls meet the minimum target size and no more. The audit tests the 24 pixels WCAG 2.5.8 requires, not the larger size most mobile guidance recommends, and many text links and small buttons on the application sit between the two. It also reads the routes it lists in the states the demo reaches, so a control that appears only in another state is not measured.

Résumé PDFs are tagged. Résumé exports to PDF carry a real structure tree (headings, lists and paragraphs a screen reader can navigate rather than one undifferentiated block of text) and declare a document language. The same résumé also exports as DOCX and plain text. Every design in the library is structurally a single reading column with real text and real headings: no tables, text boxes, or images. The cohort report PDF we send to sponsoring institutions is not yet tagged: it has no structure tree, and our conformance report says so.

No third-party audit has been performed yet. The measurements above are our own instrumentation. An independent audit producing a VPAT/ACR sits on our institutional readiness roadmap, triggered by the first engagement that requires one, and until it exists we will not imply otherwise. Our own self-assessed conformance report (VPAT, WCAG 2.1 edition), with its known gaps, is available on request.

Tell us what is in your way

If something here does not work with your screen reader, switch access, keyboard, or anything else, write to hello@standingproof.com with the page and what happened. Reports like that get fixed as defects, not filed as feedback.