01
Grip force
How hard the hand closes on the object, tip by tip, as the grasp forms.
Tactile capture for embodied AI
Nerve Instruments
Robots learn to see. They still cannot learn to feel.
A data glove that records grip pressure, palm contact, joint bend, and slip for robot-training datasets.
Fabric covered
Sensor assembly
Drag to compare ↔
The gap
Video teaches sight. It does not teach touch.
Hands show up in the frame. Contact force and slip almost never do.
Bend-only gloves
Pose without contact
Nerve Instruments
Pressure, slip, bend. Built for collection days.
Research kits
Five figures, quote only
The instrument
What the stream carries.
Wear it through a normal demo. You get a timestamped tactile track ready to sit next to your cameras, without guessing contact from pixels alone.
02
Contact distribution
Where load lands across the hand, not only at the fingertips.
03
Hand shape
How the fingers fold through the motion, including moments video cannot see.
04
Hand attitude
How the wrist turns as the hand reaches, orients, and releases.
05
Slip onset
The moment contact starts to slide, before the object visibly escapes the grasp.
Next
If you collect hand demos, let’s put touch on the same day.
Labs, data aggregators, and video-capture partners who want a tactile channel in the same session.