Talk. It responds in motion.
SolfaDuet 1 generates real-time body gestures from speech.
Live demo is currently unavailable.
Powered by SolfaDuet 1. Trained with Meta Seamless Interaction data (CC BY-NC 4.0); derived weights remain subject to the same non-commercial terms.
Current setup
SolfaDuet 1 is audio-driven in this demo; visual perception is not enabled.