Uthana builds models that create motion, data that trains motion systems, and embodiment technology that transfers movement across characters and robots.
Human motion is a missing layer in embodied AI. A character or robot needs more than geometry and appearance. It needs movement that reflects intent, responds to control, and fits the body performing it.
That motion does not transfer between bodies for free. Skeletons, proportions, joints, constraints, and output conventions change from one embodiment to another. Uthana builds the full path: generate or capture the movement, represent it as structured 3D data, and transfer it to the target character or robot.
Uthana began in games and 3D production, where motion has to work on real characters, inside real tools, under production constraints. That required more than a model demo. It required retargeting across custom rigs, editable keyframes, standard file formats, APIs, and outputs that studios could continue refining.
Physical AI extends the same underlying problem. The target body changes, but the need for structured motion, reliable transfer, and usable outputs remains. Gaming is evidence that the infrastructure works in demanding production pipelines—not a separate chapter we left behind.
We treat motion as one system from source to embodiment. Model output is not finished until it can be inspected, retargeted, exported, or delivered in the representation the next system expects.
We separate demonstrated capability from roadmap. Technical specifications, validation criteria, and unresolved constraints are defined before an enterprise engagement scales. The goal is not a broad claim about intelligence. It is motion that works in the customer's pipeline.
Uthana brings together machine learning, character animation, 3D infrastructure, and product engineering.
Viren Tellis — Co-founder and CEO
Kethan Tellis — Co-founder and CTO
Start generating motion today, or work with us on a custom dataset, embodiment system, or simulation program.
