Screening Summary
-
Clinic Terminal
Run the live interview. Upload the full session.
This is the mirror or assessment interaction surface. The realtime model handles the guided patient conversation, then the browser uploads the full recording to the batch multimodal pipeline for the formal provider-facing assessment.
Live Session
Structured realtime interview
Qwen conducts the patient interaction while the browser records the full webcam and microphone stream for later multimodal review.
Session Status
Loading realtime status...
Identity CheckNot checked
We verify the opening face before the realtime interview starts.
Live Transcript
Realtime transcript is collected as session metadata and uploaded with the full recording.
Frames Sent
0
Speech Seconds
0.0
Audio Chunks
0
Manual Upload
Analyze existing video
Keep a separate entry point for previously recorded clinic visits or externally captured patient videos.
Upload Status
Upload a video clip to run the same batch multimodal analysis without a live session.
Formal Output
Batch multimodal assessment
Final dementia-risk output is generated only after the full uploaded session is analyzed by the provider mesh.
Complete a live session or upload an existing video to render the final assessment.
Batch Provider Mesh
Upload analysis routing
This provider stack is used only for the final uploaded-video assessment, not for the realtime conversation loop.