Dexterity doesn't transfer.
Robots learn one demonstration at a time. Nothing scales across tasks, embodiments, or operators.
01 / 03
CHIROS · Research Edition · Göttingen
The instrument for human dexterity capture.
CHIROS is a data-embodying instrument that captures human dexterity in physical-AI-ready language.
The problem
Three gaps that keep Physical AI from scaling.
Teleoperation isn't a foundation.
It captures how you drive a robot, not how you perform a task. The embodiment becomes the ceiling.
02 / 03
Vision misses the physics.
No pressure. No timing. No intent. Pixels describe a scene; they don't describe a grasp.
03 / 03
Principle 01
A hand, rendered as something technology can learn from.
Every joint angle is ground truth, not an estimate.
Why Digity exists
Transferring the intelligence in human action to technology.
That's our vision. Not a slogan we reverse-engineered, the thing we embarked on in 2019, long before the humanoid wave, long before Physical AI had a name.
We saw early that this transfer needed a shared instrument. First, a tool that could acknowledge the human dimension: led by biomechatronics, ergonomics, and the way a hand actually moves under load. At the same time, it needed to be native to the technology dimension: deterministic, absolute, reproducible; legible to robots, to AI, to pattern-finding systems that need to locate themselves in data.
CHIROS is that instrument. We designed it for our own work in dexterity research, and we're releasing it because the transfer is too important, and too hard, for any single team to do alone.
01
Honour the human dimension.
Biomechatronics, ergonomics, and years on the factory floor shaped every mechanical choice.
02
Be native to the technology dimension.
Absolute encoders, deterministic timing, rigid kinematics, so the data is legible to robots and to learning systems without translation.
03
Dexterity is a shared frontier.
No single company or AI/robotics stack will transfer human dexterity to technology alone. We release CHIROS so the field can push this frontier together, as an open community.
On the workbench
Captured where work actually happens.
Under construction
Imagery from real capture sessions coming in May 2026.
Backed by
Partners, grants, and accelerators across EU and US.
NVIDIA Inception Program
EXIST Forschungstransfer
University of Göttingen
German Accelerator 2025
High-Tech GF
Ready to capture the physics of a hand?
Built in the EU. One-time purchase. Owned forever.
Research Edition · First units shipping late summer 2026