# OpenVINO™ Physical AI
OpenVINO™ Physical AI is a runtime that unifies cameras, robots, and policy inference behind a single API, accelerating OpenVINO™-powered physical-AI deployments.

It closes the perception-to-action loop at runtime, turning sensor and robot-state observations into actions with low latency for learned manipulation and task execution on real hardware.

Explore the source, documentation, and installation instructions:

::::{grid} 1

:::{grid-item-card} **OpenVINO™ Physical AI on GitHub**
:link: https://github.com/openvinotoolkit/physicalai
:link-type: url
:link-alt: clickable cards

Explore the OpenVINO™ Physical AI repository on GitHub for source code, documentation, installation instructions, and release packages.
:::
::::