2FC Physical AI Gripper

The industry is investing in vision, demonstrations and simulation, yet manipulation depends on touch. Most robots cannot feel what they grasp, so pressure, slip, force and position go uncaptured.

Tactile sensing is not new; turning raw signal into training data and real-time behavior is the hard part. The 2FC captures contact-rich data, synchronized and labeled with physical events, and runs embedded intelligence on the gripper so robots act on what they feel.

• Modalities: pressure, vibration, proprioception, force, speed, position
• Millisecond synchronization: All signals are time stamped with millisecond-level precision within a shared time reference frame.
• Unified API: Access all data through a single framework

• Grasp quality assessment: 
  "How confident am I in this grasp?"
  
  • One confidence score the cell can act on.
    
    
• Incipient slip detection: 
  "The object is starting to slip."
  
  • Automatic slip detection and labeling.
    
    
• Object pose estimation:
  "Where is the object in my fingers?"
  
  • State information for learning and downstream manipulation.
    
    
• Object centering: 
  "Move the robot to the right position."
  
  • Closed-loop manipulation.
    
    
• Advanced grip check: 
  "Did I pick something, and is it the right thing?"
  
  • Outcome verification and automatic episode labeling.

• Rated for 2M cycles: One gripper from data collection to fleet deployment.
• Compliance: Enables your system to better react, recover, and adapt to uncertainty.
• Mechanical intelligence: Patented underactuated design for both encompassing and parallel grasps.