Accessibility
Voice is optional and privacy-aware. Text, keyboard, and visible structure remain first-class. No autoplay. No information conveyed by animation or color alone.
Semantic structure, visible focus, and clear labels on controls and CTAs so the public OS pages remain usable without a mouse. Headings, lists, and landmarks should describe the page the way a reader — or assistive technology — expects. Keyboard paths are part of the design, not a retrofit after a mouse-first mock.
Where media or narration exists, transcript text is visible and complete — not hidden behind a “for accessibility” toggle. Meet Luma always exposes transcript text whether or not optional narration plays. Audio is never the only way to get the content; text remains the source of truth.
Motion respects prefers-reduced-motion. When a user asks for less motion, animated marks and decorative movement yield to static mark plus text states. Motion may create presence for people who want it; it must not be required to understand status, hierarchy, or meaning on the page.
Readable contrast targets for body copy, controls, and status labels against OS surfaces. Meaningful visuals get alternative text; decorative marks stay empty or omitted from the accessibility tree. Color alone never carries status — status labels and plain language stay available without relying on a hue.
No autoplay anywhere on the public OS experience. When narration is offered, pause, replay, and mute affordances travel with it, and starting audio is always a deliberate user action. Shared spaces and quiet clinics are assumed defaults — voice is invited, never imposed.
OneSource OS workspace · sample information
Text-first controls
Visible narration controls, readable status, and a transcript keep the experience useful without audio.
Sample information is used here; no client records are shown.
Public experience direction
Optional introduction may use browser narration with a full transcript always available beside or under the spoken path. Voice is a convenience layer for people who choose it — not a second product identity and not a substitute for keyboard and text. Shared-space safe example: “You have three items requiring attention.” Unsafe: speaking patient names, member IDs, diagnoses, claim amounts, or anything that would identify a person in a hallway or open office. If content is not safe to say aloud in a shared space, it should not be narrated by default.
Intended in-product modes include Off, Text only, Push-to-talk, Headset only, and tenant-defined restricted voice so organizations can match clinic reality. Those modes describe product direction for governed in-product use; they are not claimed as generally available on this public site. Until a mode is shipped and labeled, assume Off / text-first on public pages, with user-triggered narration only where explicitly offered.
Describe the keyboard, motion, narration, transcript, or assistive-technology consideration that matters to you.
