Introduction

The learning layer for Physical AI — turn every robot experience into better intelligence

Welcome to Humaan

Humaan is the learning layer for Physical AI—a system that connects robot experience, training data, and evaluation to drive measurable improvements in robot capabilities.

Our goal is to help developers capture what robots observe and do, understand the outcomes, and turn those experiences into data for training and evaluation. Improved models can then return to robots for testing in the real world—creating a continuous learning loop.

What you'll learn

Why Humaan exists — The mission and strategy behind building an open learning layer for Physical AI.

What Humaan is — How Humaan connects robot experience, training data, and evaluation in one continuous cycle.

Humaan Loop — The improvement cycle for robot intelligence: capture experience, prepare data, improve intelligence, and deploy and repeat.

Humaan Index — The data foundation for exploring experiences, organizing data, and curating learning-ready datasets.

Accessible Experimentation — Modular hardware and open-source software that make robot learning accessible.

Who this is for

This documentation is built for makers, roboticists, and developers who want to create systems that perceive, reason, and act in the physical world. You don't need a massive research lab — just a problem worth solving and the curiosity to turn real-world robot experience into better intelligence.

How to use this documentation

Start anywhere. If you want context on why Humaan exists, begin with the mission and strategy. If you want to understand the platform and core components, explore what Humaan is and how the learning loop works.

Intelligence is learned through experience. This documentation follows the same principle.