InfiniteTalk

Audio-driven, sparse-frame video dubbing that turns images or footage into identity-stable talking video. A streaming generator scales to effectively infinite sequences, syncing lips, head motion, posture, and expressions to speech while preserving the original camera feel.