PHANES AI and Tinno Mobile launch SeePhys Hub for full-body robot training data
The AMW Read
The launch adds a concrete wearable data product to the robotics player map, but its impact remains limited to a specialized workflow without comparative robot-learning results.
PHANES AI and Tinno Mobile launch SeePhys Hub for full-body robot training data
PHANES AI and Tinno Mobile unveiled SeePhys Hub on October 14, a wearable system that synchronizes first-person vision, hand poses, tactile pressure and full-body motion for robot learning. It combines a four-camera headband, two wristbands, three waist-and-leg trackers and tactile gloves. PHANES says the headband weighs no more than 220 grams and each wristband no more than 75 grams, with both device types supporting 24 hours of continuous capture. Its accompanying software cleans recordings, segments actions, adds semantic labels and maps human movements onto robot bodies.
The launch targets a bottleneck in physical AI: portable, vision-led human recordings can miss the body movements and contact forces needed for complex manipulation. PHANES is positioning synchronized tactile and motion capture as a way to increase learning value per recorded hour. That places it in the robotics ecosystem's data tooling layer, while supporting its stated ambition to build world models for whole-body mobile manipulation. The article provides no comparative training results showing that the additional signals improve robot performance or reduce data requirements.
For builders and investors, the concrete diligence question is whether the richer recordings produce transferable robot skills at a workable collection and processing cost. A useful evaluation would compare policies trained with and without tactile and full-body inputs on the same tasks, including the effort required for motion retargeting. SeePhys Hub establishes a product offering; its learning-efficiency advantage remains to be demonstrated.