Skip to content
I.

The Platform built by Inheritance.

Witness is the motion infrastructure built by Inheritance. It converts existing video into structured motion data that robots and world models can train on at scale. As more video is processed, the system compounds—expanding generalization across environments, embodiments, and edge cases.

II.

What It Unlocks

Video becomes durable training data. Human pose. Object interaction. Contact. Action. Context. Robotics and world model teams can train on real-world behavior without recreating it from scratch. As the corpus grows, learning accelerates.

III.

Why It Matters

Physical AI is bottlenecked by motion data, not architecture. The systems that define this layer determine how quickly robots learn. Witness defines that layer.