Input Audio
Source audio provided: Nest health mock visit.m4a
Bundled demo audio is loaded from this project so playback is consistent for the Aiden and Sebastian scenario.
FAMILY VISIT: Recording Attribution Demo
Family Visit Speech Attribution
This version auto-generates a patient-attributed event stream and patient-specific clinical findings from transcript text using multi-signal attribution checks for each finding (explicit name, context continuity, and speaker cues).
Pipeline View
1. Ambient Capture chunk audio into 5-10 second windows
2. ASR + Diarization transcript text plus speaker signals
3. Patient Attribution explicit mention, speaker identity, context, switch detection
4. Clinical Extraction symptoms, vitals, exam, and plan per patient
Transcript to Parse
Edit transcript text below, then click Generate Attribution.
Generated Event Stream
Format: timestamp | speaker | patient | text | confidence
Row-level editing is disabled for this simplified demo. Clinician edits happen in the patient findings section.
Attributed Events
0Identified Patients
0Unassigned Events
0Multi-Patient Events
0| Timestamp | Speaker | Patient | Transcript | Confidence |
|---|
Clinician Actions
Update findings as needed, then submit a demo payload to Athena.