NVIDIA’s ovrtx library has taken another step toward bringing gaussian splatting assets deeper into Omniverse based simulation workflows. In its 0.2.0 release, the lightweight C and Python SDK added support for Gaussian Splats and other particle field primitive types through the UsdVol.ParticleField schema, extending a library NVIDIA positions around physically accurate, real time sensor simulation and visualization for robotics, synthetic data, and industrial workflows.
This ties NVIDIA’s renderer more directly to the same standardization wave now moving through OpenUSD. The ParticleField schema is designed as a base for particle based volumetric primitives, and OpenUSD’s ParticleField3DGaussianSplat schema specifically describes the original 3D Gaussian Splatting representation, including built in attributes and rendering hints for splat data.
The release also broadens ovrtx in ways that should appeal to developers building more serious applications on top of it. NVIDIA added an operation status query API and logging callback for monitoring renderer operations, new helper functions for creating supported config values, CUDA GPU selection controls at both renderer creation and render product level, and Python side version checks to confirm compatibility between the installed package and native library. Python bindings were also expanded for profiling controls, GPU transform reads, and asynchronous attribute writes, bringing the Python experience closer to the underlying C API.
ovrtx is still labeled pre-release software, and the repository only launched publicly in February 2026, but NVIDIA is already moving it beyond camera only basics toward a broader scene representation story inside Omniverse. With gaussian splats now entering through USD particle field schemas, the library begins to look like another building block in NVIDIA’s larger push to make real world captures, sensor simulation, and photoreal digital environments work together inside physical AI workflows.
ovrtx comes with the standard NVIDIA license and can be found here on GitHub.
