Realtime Browser Demo

Azure Speech Streaming + Sana Supervisor

Continuous Azure Speech recognition stays live while Sana keeps the explicit PRO-2 state machine. This path is meant to feel materially faster than the upload/transcribe loop and includes the named greeting and final health-share step.

Evidence-based scope

Guided interview limited to PRO-2 plus one final health-share question.

Safety boundary

Data collection only. No treatment advice, no off-protocol symptom expansion.

Progress rule

Remaining required items: 0

Profile

SDK: loading · Recognition: idle · Playback: idle

InterviewerParticipant

Conversation Log

No turns yet.

Azure Stream Events

Recognizer, interruption, synthesis, and finalize events for the Azure Speech path.

No Azure events yet.