Swing Choreo debuts: multi-person text-to-motion model that follows your beats
andrew_n_carr · x · 2026-10-14
Andrew Carr introduced Swing Choreo, a new human motion model. Describe a performance one beat at a time and Choreo performs all of it: each action in order, hands and faces, several people at once, moving where you tell them to.
Mitchell Probst notes Carr shared the dream of a multi-person text-to-motion model years ago—what seemed far away then is now available for everyone to create with.
Related event: Swing Releases Choreo, a Beat-by-Beat Human Motion Generation Model(10 posts)→
More from Multimodal
- Building Blender scenes entirely with AI: daytime lighting and fog handled far better — jongranskog · 2026-10-14
- AI-generated "Fallout 5" episode goes viral with its Rated-R TikTok-style opener — Dear_Hovercraft_6500 · 2026-10-14
- Moondream 3.1 and Clef Go Live on Prime Inference and OpenRouter — xeophon · 2026-10-14
- Claude rebuilt my piano 3D VFX pipeline that once took Unity and manual camera calibration — tracyhenry400 · 2026-10-14
- ComfyUI seed-hunting nodes for MiniMax H3 video generation coming to GitHub soon — smereces · 2026-10-14
- Fish Audio's open-weights Drama 3 Preview tops TTS leaderboards, beats Eleven v4 on pronunciation — ArtificialAnlys · 2026-10-14