Physically accurate assets for
Gazebo
Convert 3D models, images, or text into assets with accurate mass, inertia, friction, and collision geometry ready for Gazebo and ROS 2 simulation pipelines.
The problem
Why existing workflows fall short.
SDF/URDF physics is tedious
Gazebo models need accurate inertial properties, collision geometry, and friction parameters. Estimating these manually for each object is error-prone.
Limited model libraries
Gazebo model repositories have limited variety. Custom environments need hundreds of unique objects with realistic physics.
ROS 2 pipeline friction
Getting 3D assets from design tools into a ROS 2 simulation pipeline involves multiple conversion steps and manual physics annotation.
How Rigyd helps
AI-native infrastructure that automates the hard parts.
Automatic inertial estimation
Rigyd estimates mass, center of mass, and inertia tensors using AI-powered material identification and geometry analysis.
Collision mesh generation
Convex decomposition generates collision geometry that balances simulation accuracy with performance, configurable for navigation or manipulation tasks.
OpenUSD interoperability
Rigyd outputs OpenUSD and MJCF natively. For Gazebo, the OpenUSD output can be imported via Gazebo Sim's USD support or converted to SDF/URDF using community tools, maintaining physics properties across format boundaries.
cost reduction in asset preparation
better real-world performance with physically accurate training data
from upload to simulation-ready asset
Accelerate your Gazebo simulations
Drop in a 3D model, image, or text description and get physically accurate assets for ROS 2 in minutes.
Frequently asked questions
Does Rigyd output SDF and URDF directly?
Rigyd outputs OpenUSD and MJCF directly. SDF and URDF for Gazebo and ROS 2 pipelines are produced by converting the OpenUSD output using community USD→URDF/SDF tools; physics properties (inertia, friction, collision geometry) are preserved across that conversion without manual re-annotation.
How accurate are the estimated inertial properties?
Mass lands within 15-20% of measured values and friction within 0.1 coefficient, inside the variance range typical domain randomization targets. Accuracy is calibrated against NeRF2Physics (CVPR 2024) benchmarks.
Can I use Rigyd for ROS 2 simulation pipelines?
Yes. Generate SimReady OpenUSD once with Rigyd, then either load it directly via Gazebo Sim's USD imports or convert to SDF/URDF with community tools for legacy Gazebo. Enterprise API integrates into CI pipelines so new assets enter simulation as they're added to the catalog.
Does Rigyd output SDF or URDF directly, or only OpenUSD?
OpenUSD and MJCF are the canonical native outputs. SDF and URDF are produced by converting the OpenUSD output using community USD→URDF/SDF tools; the advantage of authoring once in USD is that the derived SDF/URDF preserves mass, inertia, and collision-mesh data. The resulting XML is well-formed and passes Gazebo's built-in model validator.
How does Rigyd handle Gazebo contact and friction solver tuning?
Friction coefficients use Gazebo's ODE solver conventions (mu, mu2 for anisotropic surfaces) populated from material identification. Contact softness (kp, kd) defaults to physically plausible values per material class. Override via the Rigyd dashboard if your scenario needs hand-tuned contact (e.g. specific compliance for soft-finger grippers). For projects using the newer DART or Bullet physics backends in Gazebo Garden and beyond, the same friction values translate cleanly, solver-specific tuning parameters can be overridden separately so the same asset works across the Gazebo Classic and Ignition/Gazebo Modern stacks.
Related reading
In-depth guides on robotics simulation, OpenUSD, and SimReady assets.
9 JUL 2026
Scaling Simulation Asset Libraries Beyond Curated Inventory
Curated SimReady libraries are a great starting point but a hard ceiling. Here is why fixed inventories limit robotics simulation at scale, and how on-demand asset generation closes the gap.
2 JUL 2026
Domain Randomization for Robotics Training: Asset Diversity at Scale
Domain randomization makes trained policies transfer to the real world, but it only works if your asset library is diverse enough. Here is how asset diversity drives randomization, and how to generate it at the scale training needs.
30 JUN 2026
Building a Scalable Embodied AI Asset Pipeline: From Raw Data to Simulation
A practical look at the stages of an embodied AI asset pipeline, from raw 3D data and reference inputs to physics-ready simulation assets, and what changes when you need to produce thousands of them instead of a handful.
Explore more
SimReady for Isaac Sim
OpenUSD assets validated for NVIDIA Isaac Sim
Models for MuJoCo
Physically accurate assets for robot learning
Assets for Gazebo
Simulation-ready models for ROS 2
Warehouse Simulation
50,000+ SKUs need physics properties
Humanoid Robots
Whole-body interaction training assets
Sim-to-Real Transfer
Close the gap with accurate physics data