OpenUSD for Developers
Developed by Pixar Animation Studios, OpenUSD is an open-source framework for creating, simulating, and collaborating in 3D worlds. OpenUSD is foundational to NVIDIA Omniverse™, the platform for developing 3D applications for industrial digitalization and generative physical AI.
How It Works
OpenUSD is an open-source framework for creating, simulating, and collaborating in 3D worlds. Developed by Pixar Animation Studios, OpenUSD goes beyond a file format to support 3D content creation and tool interchange.
The universal structure of OpenUSD allows developers to integrate multiple assets into a single "stage," enabling dynamic scene manipulation. Each OpenUSD asset contains independent data layer stacks, such as geometry, shading, or textures, which can be interchanged without affecting other elements. For example, in an OpenUSD stage in a kitchen environment, the data layers for a chair or stove can be swapped independently, enabling rapid scene updates.
This dynamic manipulation empowers developers to efficiently test multiple asset versions based on context, streamlining scene creation and updates without the need to rebuild from scratch. The versatility of OpenUSD has established it as an industry standard across fields like visual effects, architecture, robotics, manufacturing, and more.
OpenUSD Overview
Check out this guide for samples, technical references, and tools to help you get started with OpenUSD.
How to Use OpenUSD Blog
Learn how developers are adopting OpenUSD in the era of AI and industrial digitalization to accelerate their 3D workflows.
OpenUSD Explainer Video
Watch this video for an overview of OpenUSD and its role in accelerating 3D development for the next era of industrial and physical AI.
Learn OpenUSD
Curriculum
Accelerate your OpenUSD workflows with this free curriculum for developers and 3D practitioners.
Ways to Get Started With OpenUSD
Use pre-built libraries to start exploring and learning with OpenUSD content like USD samples. You can also start developing with OpenUSD using the latest USD Python build or API.
Download Pre-Built OpenUSD Libraries and Tools
Download a full build of OpenUSD, which includes pre-built libraries and the USD toolset. The Windows download also comes fully bootstrapped so you can immediately execute usdview or any other bundled USD tool.
By downloading or using the pre-built OpenUSD libraries, you agree to the terms of the NVIDIA Omniverse License Agreement.
Download the usd-core Python Package on PyPI
The usd-core package is also available if you want to experiment with the USD Python API or if you only need the core USD features for reading and writing USD stages and layers. USD comes pre-built, and it's just a "pip install" away.
OpenUSD Example Projects and Sample Assets
Free and open-source USD samples are available for download through NVIDIA Omniverse and partner sites. Explore USD example datasets, physical AI datasets, and the NVIDIA Learning Library for more OpenUSD resources.
OpenUSD SimReady Standardization Workflow
Standardized requirements and processes to enable development of simulation-ready capabilities and assets for AI factories.
Physical AI Dataset
This commercial-grade, pre-validated dataset can help researchers and developers pretrain, test, and validate AI models for physical AI deployment.
Da Vinci Workshop Project
The da Vinci Workshop showcases how the NVIDIA creative team structured this amazing environment to develop a complete USD film production pipeline.
Residential Lobby Project
The Residential Lobby Project showcases the use of multiple USD composition arcs to help a multi-person design team create an architectural environment.
Factory Planning 3D Sample Asset
USD Explorer’s sample 3D model uses various CAD applications to show a sample factory model and factory planning.
Warehouse Sample Asset
This industrial 3D model shows a sample warehouse space using Pixar USD models of bins, crates, safety cones, barrels, conveyors, ramps, ladders, and pallets.
Built on OpenUSD Technology
Leverage OpenUSD technologies for robotics and industrial simulation workflows.