Physical AI for MedTech is here — where intelligent systems sense, perceive, and act in real time. Devices are born in simulation, learn new skills with Isaac for Healthcare, and deploy on NVIDIA Holoscan for multi-sensor, multi-modal applications at the edge.
This organization is the home of Clara Medical Models — open foundation models to segment, reason, and generate across medical data. The model weights are distributed through hugging face Clara medical collection. Built with MONAI, an NVIDIA co-founded open source toolkit for Medical AI research & development.
| NV-Segment | NV-Generate | NV-Reason |
| GitHub • HF-CT • HF-CTMR | GitHub • HF-CT • HF-MR | GitHub • HuggingFace |
 |
 |
 |
| Segments anything in 3D CT & MR volumetric images across 345+ anatomical structures through automatic detection & interactive point-click refinement workflows. | Generates high-resolution synthetic 3D CT & MR volumetric images with anatomical annotations to augment medical datasets while preserving patient privacy. | A vision-language model that uses chain-of-thought reasoning to express medical thinking step-by-step, based on how radiologists analyze chest X-rays. Designed to advance explainable AI research and bring transparency to Medical AI. |