Clinical Scribe
AI-assisted behavioral health documentation — signal-informed, clinician-signed.
Therapist · DAP
MD / Psychiatrist
Case Manager
Sarah M.
37F · Day 31 · IOP · OUD/AUD
Microsoft Azure Kinect · FACS Analysis
Facial Action Coding System (Ekman) — 26 muscle-level AU codes · 4 kinematic signals · session active
LIVE
Voice analysis · C-SSRS proxy
—
Session Capture
0:00
Confirm consent above to enable
Please confirm patient consent above
Dr. Rivera transcribing session…
or
Switch to paste mode →
CLINICIAN: [text] CLIENT: [text]
Synthetic transcript loaded
Generating in 3
⚡
Sarah M.'s signals are loaded
247 signals · Therapist · DAP · Demo only
Prior note (optional — enables compression comparison)
Upload prior note
.txt · or drag and drop · or paste below
The Council synthesizing signals…
Compression scorecard
Compression
—
vs transcript
Council synthesized — of 374 signals this session —
Wearable (12 devices): HRV, resting HR, SpO₂, skin conductance, sleep stages, activity cadence ·
Video (30): 26 Ekman FACS action units + 4 body kinematics via Azure Kinect ·
Voice (240): C-SSRS–anchored phrase library — hopelessness, ideation, withdrawal, affect flattening, agitation ·
Behavioral (92): app engagement, session attendance, sleep regularity, social isolation index, medication adherence
Generated note — Dr. Rivera
72%
Auto-filled from patient signals
Green sections trusted · Amber sections verify · Red sections require your input
3 sections need your input
Signal data — auto-trusted
Verify & edit recommended
By signing you confirm this note accurately reflects the clinical encounter and the signal data above. Time-to-sign is logged for quality metrics.
Beta feedback
Was this note shorter than what you'd typically write?
Did the wearable and FACS signals add clinical value?
Was your formatting preserved?
Anything else?
Feedback recorded — this shapes the next build.
