Musetalk

For five weeks we would not say how our presenter’s talking footage is made. Here is the whole chain: LongCat for the body, MuseTalk for the mouth, Duix to clean up what that leaves, and MuseTalk again for the visemes. Why the order matters more than any single model.

Our character’s debut music video, 18 shots, five sung sections, 2:13, rendered overnight on a single RTX 5090 for zero cloud spend. The complete pipeline: song, storyboard, renders, lip-synced vocals, beat-cut assembly.

MuseTalk generates lip-synced video faster than real time on a 5090, but getting it to BUILD on Blackwell is dependency hell. Here’s the exact recipe that works.