ovphysx is a self-contained library for USD-based physics simulation, offering a C API with Python bindings. It wraps NVIDIA PhysX and Omni PhysX, loads USD scenes, runs simulation, and exchanges data via DLPack tensors — no Omniverse installation required.

Note: Pre-release notice: ovphysx is pre-release software and not yet mature. Current limitations include a strict in-process USD requirement: ovphysx can only coexist with OpenUSD v25.11 that is non-monolithic, Python-enabled, and linked against oneTBB (in particular, ovphysx cannot be used together with usd-core in the same process today), and parts of the API are still being completed and may change before 1.0.

pip install ovphysx

from ovphysx import PhysX

physx = PhysX()
physx.add_usd("scene.usda")
physx.step(1.0 / 60.0, 0.0)
physx.release()

Clone environments for batched reinforcement-learning workloads:

from ovphysx import PhysX
from ovphysx.types import TensorType
import numpy as np

physx = PhysX(device="cpu")
usd_handle, _ = physx.add_usd("scene.usda")
physx.wait_all()

# Clone env0 to create 3 additional environments
physx.clone("/World/envs/env0", ["/World/envs/env1", "/World/envs/env2", "/World/envs/env3"])
physx.wait_all()

# Read rigid body poses across all environments via DLPack-compatible tensors
pose_binding = physx.create_tensor_binding(
    pattern="/World/envs/env*/table",
    tensor_type=TensorType.RIGID_BODY_POSE,
)
poses = np.zeros(pose_binding.shape, dtype=np.float32)
pose_binding.read(poses)

pose_binding.destroy()
physx.remove_usd(usd_handle)
physx.release()

A standalone C SDK is available for integration into non-Python applications. Download pre-built packages from the GitHub Releases page.

After extracting the SDK, you can build and run a bundled sample directly:

# /path/to/ovphysx-sdk is the extracted SDK package (pre-built binaries from GitHub Releases)
cmake -B build -S /path/to/ovphysx-sdk/samples/c_samples/hello_world_c -DCMAKE_PREFIX_PATH=/path/to/ovphysx-sdk
cmake --build build
./build/hello_world_c

The SDK includes ready-to-build samples in samples/c_samples/ covering core workflows (hello world, tensor bindings, cloning, contacts, OmniPVD recording). Each sample has its own CMakeLists.txt that uses find_package(ovphysx).

• User Guide & Tutorials
• API Reference
• Samples

• Python 3.10+
• Linux (x86_64, aarch64) or Windows (x86_64)
• NVIDIA GPU + driver recommended (CPU-only simulation also supported)

The ovphysx source code is licensed under the BSD-3-Clause License — see LICENSE.txt.

Pre-built binary distributions (SDK packages and Python wheels) are licensed under the NVIDIA Omniverse License.

• PyPI
• Documentation
• Issues
• Discord