Accessibility
Accessible by default, not by afterthought.
Election observation has to work for the person using it, not just the person watching a demo. This is a plain account of what we've actually tested, what it found, and, just as importantly, what it doesn't yet cover.
What we've tested
An automated accessibility scan (axe-core, driving a real browser, not a static code check) was run against every major screen in both the Custody dashboard and the field app, in every one of the 5 supported languages, in both the platform's dark and light themes. The most recent run found zero violations across all 51 screens scanned.
On top of the automated scan, a manual keyboard-only walkthrough (no mouse) was run across the core dashboard, agent management, report review, and incident-creation flows, checking every stop for reachability and a visible focus indicator. One real, reproducible gap was found this way: several of the field app's primary buttons had no visible focus ring on keyboard focus. It was fixed and confirmed with a real Tab keypress, not just a visual check.
What this doesn't claim
Zero automated violations is a real result, not a rounding error, but it's a statement about one tool's ruleset, not a certification of WCAG 2.1 AA compliance. Automated tools including the one used here are widely estimated to catch only a third to half of real accessibility issues. We'd rather tell you that plainly than let a clean scan imply more than it proves.
What's still outside this pass
This is a living list, not a closed ticket. The next pass should start from item one below, not from zero.
Screen-reader testing with an actual screen reader (VoiceOver, NVDA, JAWS, TalkBack). Everything to date was verified through automated DOM analysis, not by listening to a real screen reader read these screens aloud. This cannot be meaningfully automated, it needs a human tester.
Cognitive and plain-language review of form error messages, empty states, and instructional copy, not yet evaluated for reading level or clarity, which matters for agents working under stress in the field.
A full manual keyboard walkthrough of every remaining screen. The flows explicitly walked through so far are covered, several admin tabs and other screens have only passed automated focus-order checks, not a human tabbing through them.
Whether status and severity indicators, like incident severity or verification-status badges, convey their meaning through more than color alone, for colorblind users. Several already pair color with a text label, but this hasn't been checked systematically.
Text scaling at the mobile viewport width specifically, as opposed to browser zoom, which has already been tested and fixed at 200%.
Report an accessibility issue
If something on this platform doesn't work for you, a screen reader, keyboard navigation, color contrast, or anything else, tell us directly. We treat an accessibility report the same way we treat a security report: seriously, and without asking you to prove it first.
Talk to us