Object detection and audio transcription running simultaneously through the AI router. Video frames are sent to YOLO, audio chunks to Whisper — each routed to the best available GPU node.
STATUS
Idle
DETECTIONS
0
TRANSCRIPTIONS
0
YOLO Latencydetect--ms
MIN -AVG -MAX -SAMPLES 0
Whisper Latencytranscribe--ms
MIN -AVG -MAX -SAMPLES 0
Requires camera & microphone access
Object Detection YOLO
Press Start to enable the camera
Audio Transcription Whisper
Transcript will appear here
Activity Log
Text embeddings -- the fourth modality alongside vision, speech, and language. Tiny model, high volume, and the best fit for CPU-only capacity: every call here is cheap enough that a GPU would be overkill.