Data / Robotics Data
Data for robots that act in the physical world.
Embodied AI needs data that text cannot provide: how hands move, how objects respond, and how tasks unfold in real environments.
01 /Data types
What we source, generate, and evaluate.
01
Ego-view
First-person video of tasks performed in real settings.
02
Wrist camera
Close-range manipulation views aligned with actions.
03
UMI
Universal Manipulation Interface captures for policy learning.
04
Tactile
Touch and force signals paired with visual context.
05
Teleoperation
Operator-controlled robot trajectories and corrections.
06
Human demonstrations
Recorded demonstrations of manipulation and task sequences.
02 /How it is made
Every example carries a record of how it was made.
Source→Generate→Review→Evaluate→Exchange
This category is delivered through FelixSphere data engagements, with the same review and provenance standards used on UnifyData.
Built for
- Robotics foundation model teams
- Humanoid and manipulation companies
- Research labs
Compute · Data · Models · Intelligence
Tell us what your model needs to learn.
Describe the capability, the domain, and how you will measure it. We will scope the data.