Shipping ProHand: The Human-Like Robotic Hand Built for Dexterous Manipulation

Over the past decade, robots have become stronger, faster, and more capable. They can navigate warehouses, run marathons, and perceive the world with advanced vision systems. But most robots still rely on simple grippers to interact with the physical world, and most robotic hands sacrifice dexterity for simplicity. These systems work in structured environments, but they quickly break down when tasks require contact-rich, human-like manipulation.

Proception

Human hands, by contrast, are extraordinarily capable. With more than twenty degrees of freedom in each hand, dense tactile sensing, and tendon-driven actuation, the human hand can perform a huge range of tasks, from tying shoelaces to repairing electronics. Unlocking this level of dexterous manipulation is one of the keys to making humanoid robots useful in the real world.

Today, we are introducing ProHand 1.0, a high-dexterity robotic hand designed for researchers and robotics companies building contact-rich manipulation systems.

Hardware that mimics the human hand#

We worked closely with hand surgeons to design ProHand around key principles of human hand anatomy:

Human-like kinematics: Twenty-two degrees of freedom and multiple joints per finger enable a wide range of dexterous motions.
Tendon-driven actuation: Similar to the way muscles pull tendons in the human body, motors actuate cables that move the fingers while keeping the fingers lightweight and compact.
Compact, robust mechanical design: The system is engineered to deliver high dexterity, withstand impacts, and remain easy to service — all in a compact form factor.
Skin-like sensors: Integrated skin-like sensors detect contact and support grip control during manipulation.

ProHand hardware Proception

A data platform to train on human data#

Dexterous manipulation is not only a hardware challenge, it's also a data problem. Modern AI systems have made remarkable progress in vision and language, fueled by the abundance of large-scale training datasets. But manipulation has historically lacked the same kind of scalable, real-world data.

Most robot manipulation data today is collected through teleoperation. ProHand supports standard teleoperation workflows out of the box, but this training method has two major limitations:

It doesn't scale easily. Data collection is constrained by the number of robots available, and usually happens inside robotics labs.
It loses important human interaction signals. When a person teleoperates a robot, the operator is not directly touching the object, so the data often misses the subtle contact, pressure, and adjustment strategies humans use during manipulation.

Because our hardware so closely resembles a human hand, we've been able to build the first data platform that can use actual human data for training. ProGlove turns the same sensor skin that encloses ProHand into a wearable data collection system for human hands. It captures human hand interaction data without requiring a robot in the loop.

With ProGlove and a headset for visual context, AI researchers at the world's leading robotics labs can collect real human manipulation data simply by wearing a sensor glove and transfer those movements and skills directly to the robot.

Proception

Shipping June 29th to researchers and robotics companies#

Proception has raised $11 million in seed funding led by First Round Capital, with participation from Y Combinator and BoxGroup. This funding allows us to expand the team, scale production, and continue building the hardware and data infrastructure needed for dexterous manipulation.

Our team has designed consumer electronics at Apple that reached millions of users, built robotic systems at Tesla and early-stage startups, and now we are building ProHand right here in Palo Alto. Solving robotic manipulation will open the door to applications across manufacturing, logistics, healthcare, and home assistance.

We are excited to collaborate with researchers and robotics companies working toward the same goal. The first batch of units is shipping this week. Researchers and robotics teams can go to our website to order.

If you're interested in ProHand/ProGlove or want to join us, we would love to hear from you.