Co-Generating 3D Motion and Video for Arbitrary Rigs
October 19, 2026
This model co-generates replayable 3D motion and synchronized video for any supplied rig, rather than being limited to human skeletons. It utilizes a separate 3D motion model connected to a pretrained video model via a small separator to enable scene-registered animation.
HOW THIS AFFECTS YOU
●
builderThis architecture allows for much greater flexibility in rigging compared to human-only models.
●
designerYou can generate controllable, 3D-animatable content from video prompts.