# Rigyd > Rigyd turns a 3D file, an image, or a text prompt into a physics-enabled SimReady asset — OpenUSD and MJCF with mass, friction, collision geometry and articulation derived rather than guessed — for robotics simulation in Isaac Sim, Isaac Lab and MuJoCo. This is the index for rigyd.com, the product and editorial site. Notes for agents: - Every page here is available as Markdown. The links below point at the `.md` twin directly; the same content is served from the canonical URL with `Accept: text/markdown`. - `https://rigyd.com/openapi.json` describes the endpoints this origin serves (the demo-asset gallery and the access-request form). It is not the product API. - **The product API is documented elsewhere.** Base URL, authentication, endpoints and credit costs are in , with the full text at . Nothing about the API is restated on this site. - `https://rigyd.com/sitemap.xml` lists every canonical URL. A path without a trailing slash redirects (301) to the canonical form. ## Start here - [Rigyd](https://rigyd.com/index.md): What Rigyd does: the agent that researches an object before it builds it, a live SimReady viewer, and the questions engineers ask first. ## Comparisons Table-first comparisons of simulators, description formats and collision tooling, including where Rigyd does not fit. - [Rigyd vs NVIDIA SimReady Asset Library](https://rigyd.com/compare/rigyd-vs-nvidia-simready-library/index.md): How Rigyd compares to NVIDIA's SimReady asset library: curated free assets vs converting your own raw 3D, images, and text. Coverage, cost, physics accuracy, formats, and when to use each. - [Isaac Sim vs MuJoCo vs Gazebo](https://rigyd.com/compare/isaac-sim-vs-mujoco-vs-gazebo/index.md): A balanced comparison of NVIDIA Isaac Sim, MuJoCo, and Gazebo for robotics simulation: physics engines, rendering, GPU acceleration, asset formats, ROS support, and which to choose for your use case. - [Isaac Sim vs MuJoCo](https://rigyd.com/compare/isaac-sim-vs-mujoco/index.md): A platform and an engine, compared directly: licences (both Apache 2.0 now, but not in the same way), PhysX 5 vs MuJoCo's own solver, rendering, sensors, hardware floor, ROS 2, and where the two are converging. - [Isaac Sim vs Gazebo](https://rigyd.com/compare/isaac-sim-vs-gazebo/index.md): A platform and a set of libraries, compared directly: the Apache 2.0 seam in Isaac Sim's licence, PhysX 5 against Gazebo's pluggable engines, the NVIDIA GPU floor, sensors, formats, and why Gazebo Classic makes most older comparisons useless. - [MuJoCo vs Gazebo](https://rigyd.com/compare/mujoco-vs-gazebo/index.md): A physics engine and a robot simulator, compared directly: contact accuracy, parallel simulation, sensor support, ROS 2 integration, and asset formats. Which to pick, and why many teams run both. - [PyBullet vs MuJoCo](https://rigyd.com/compare/pybullet-vs-mujoco/index.md): Two Python-first physics engines, compared directly: Bullet's LCP solvers against MuJoCo's convex contact model, the release-cadence gap, MJX's GPU and TPU scaling, and why PyBullet reads more formats — MuJoCo's own included. - [OpenUSD vs SDF vs URDF vs MJCF](https://rigyd.com/compare/openusd-vs-sdf-urdf-vs-mjcf/index.md): A balanced comparison of the four scene and robot description formats used in robotics simulation, OpenUSD, SDF, URDF, and MJCF. Strengths, limitations, ecosystem support, and which to author in. - [Manual vs AI Physics Annotation for Simulation Assets](https://rigyd.com/compare/manual-vs-ai-physics-annotation/index.md): Hand-authoring physics for simulation assets vs AI-automated estimation: time, cost, accuracy, and scale compared. When manual annotation still wins and when automation is the only practical option. - [Rigyd vs NVIDIA USD Composer](https://rigyd.com/compare/rigyd-vs-nvidia-usd-composer/index.md): NVIDIA USD Composer (formerly Create) assembles OpenUSD scenes; Rigyd prepares the assets that go into them. How they compare on physics, asset source, and when to use each. - [URDF vs MJCF for ROS 2](https://rigyd.com/compare/urdf-vs-mjcf-for-ros-2/index.md): URDF is the ROS standard; MJCF is the MuJoCo native format. With MuJoCo + ROS 2 integrations now mainstream, which should you author in? Expressivity, tooling, and the practical recommendation. - [Isaac Lab vs Isaac Sim](https://rigyd.com/compare/isaac-lab-vs-isaac-sim/index.md): Isaac Sim is NVIDIA's robotics simulator; Isaac Lab is the unified learning framework built on top of it. How they relate, when you need each, and where OmniIsaacGymEnvs and Isaac Orbit fit. - [Isaac Lab vs MuJoCo Playground](https://rigyd.com/compare/isaac-lab-vs-mujoco-playground/index.md): Two robot-learning frameworks, compared directly: BSD-3 against Apache 2.0, PhysX against MJX, the OpenUSD and MJCF asset split, the hardware floor each assumes, and the Newton backend that is closing the gap between them. - [CoACD vs V-HACD vs manual hulls](https://rigyd.com/compare/v-hacd-vs-coacd-vs-manual-hulls/index.md): CoACD is the collision-aware convex decomposition algorithm that emits fewer hulls than V-HACD, so assets simulate faster. Speed, hull count, sim performance, and when each approach wins. - [Convex hulls vs triangle mesh colliders](https://rigyd.com/compare/convex-hulls-vs-triangle-mesh-colliders/index.md): A convex hull is a convex approximation the solver collides against cheaply; a triangle mesh collider is the exact surface, and in most engines it only works on static bodies. The static-vs-dynamic rule engine by engine, and why decomposition exists. ## Resources Long-form technical guides on physics estimation, OpenUSD and asset pipelines. - [Automatically Estimating Physics Properties for Simulation Assets](https://rigyd.com/resources/automatically-estimating-physics-properties-for-simulation-assets/index.md): Mass, friction, and collision geometry are what make a 3D model behave in a physics engine. This is how those properties get estimated automatically, and why automating them is what lets teams build simulation assets at scale. - [Scaling Simulation Asset Libraries Beyond Curated Inventory](https://rigyd.com/resources/scaling-simulation-asset-libraries-beyond-curated-inventory/index.md): 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. - [Domain Randomization for Robotics Training: Asset Diversity at Scale](https://rigyd.com/resources/domain-randomization-for-robotics-training/index.md): 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. - [Building a Scalable Embodied AI Asset Pipeline: From Raw Data to Simulation](https://rigyd.com/resources/building-a-scalable-embodied-ai-asset-pipeline/index.md): 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. - [Text and Image to Simulation Asset: Multi-Modal Input for Robotics Teams](https://rigyd.com/resources/text-and-image-to-simulation-asset-generation/index.md): How robotics and ML teams turn text descriptions, reference images, and existing 3D files into physics-ready simulation assets, and why multi-modal input matters when you are building training environments at scale. - [What is a SimReady asset and why does it matter for robot learning and evaluation](https://rigyd.com/resources/what-is-a-simready-asset-and-why-it-matters-for-robot-learning/index.md): A SimReady asset is a 3D model carrying the physics, collision, material, and semantic metadata a simulator needs to use it directly. This guide explains the anatomy of a SimReady asset and why it is the bottleneck for training and evaluating robot policies. - [How Robotics Engineers and ML Teams Building Embodied AI Systems Can Convert 3D Models to Simulation Assets](https://rigyd.com/resources/how-to-convert-3d-models-to-simulation-ready-assets/index.md): A practical guide for robotics engineers and ML teams on converting 3D CAD models into physically accurate, simulation-ready assets for training and evaluating embodied AI systems at scale. - [Rigyd vs NVIDIA SimReady Alternatives: A Guide for Robotics and ML Teams](https://rigyd.com/resources/rigyd-vs-nvidia-simready-alternatives/index.md): Compare the best alternatives to NVIDIA SimReady asset libraries for robotics engineers and ML teams building physically accurate simulation environments. - [Synthetic data generation for computer vision in robotics](https://rigyd.com/resources/synthetic-data-generation-for-computer-vision-in-robotics/index.md): Real-world labeled robotics data is expensive and slow to capture. Synthetic data is fast and unlimited, but only useful if the underlying simulation has correct physics, semantic labels, and domain randomization. Here's the complete pipeline. - [Digital twin creation pipeline for manufacturing](https://rigyd.com/resources/digital-twin-creation-pipeline-for-manufacturing/index.md): A factory digital twin needs every object to behave physically, not just render. This is the end-to-end pipeline: CAD intake, BIM merge, physics layer, semantic labeling, simulation runtime, at the asset volumes (10K+ unique SKUs) real factories actually contain. - [How to set up mass, friction, and joint properties for robot training](https://rigyd.com/resources/how-to-set-up-mass-friction-and-joint-properties-for-robot-training/index.md): The three pillars of robot physics setup, mass, friction, joints, determine whether your trained policy transfers to real hardware. Here's the calibration target for each, the schemas, and the common mistakes that quietly break training. - [Best practices for collision meshes in robotics sim](https://rigyd.com/resources/best-practices-for-collision-meshes-in-robotics-sim/index.md): The collision mesh is the single most-tuned property in robotics simulation. Wrong choice = phantom interpenetration, slow physics, or both. Here's how to pick between primitives, convex hull, V-HACD decomposition, and mesh simplification, with concrete defaults per asset class. - [How to convert GLTF/FBX/OBJ to OpenUSD for simulation](https://rigyd.com/resources/how-to-convert-gltf-fbx-obj-to-openusd-for-simulation/index.md): Most 3D pipelines export to GLTF, FBX, or OBJ. Most simulators expect OpenUSD. This guide covers the conversion paths, Omniverse, Blender, command-line, and AI-automated, plus what each format preserves and what gets lost. - [Sim-to-real transfer: why physics accuracy matters more than visual fidelity](https://rigyd.com/resources/sim-to-real-transfer-physics-accuracy/index.md): The sim-to-real gap is overwhelmingly a physics problem, not a rendering problem. Here's the research, the failure modes, the calibration ranges that actually matter, and why investing in physics accuracy beats investing in photorealism for most robotics policies. - [Isaac Sim asset requirements and best practices](https://rigyd.com/resources/isaac-sim-asset-requirements-and-best-practices/index.md): NVIDIA Isaac Sim expects assets in a specific format: OpenUSD with USDPhysics schemas, calibrated mass and inertia, convex collision meshes, semantic labels, and validated material bindings. This is the practical checklist. - [How to create SimReady assets without manual modeling](https://rigyd.com/resources/how-to-create-simready-assets-without-manual-modeling/index.md): Building a SimReady asset by hand takes ~4 hours and requires Blender, V-HACD, and USDPhysics expertise. AI-driven asset preparation collapses the workflow to about 5 minutes per asset. Here's how the automated pipeline works and when to use it. - [What is OpenUSD and why does it matter for robotics](https://rigyd.com/resources/what-is-openusd-and-why-it-matters-for-robotics/index.md): OpenUSD is the 3D scene description format Pixar built for film and NVIDIA scaled for robotics. This guide explains the architecture, the USDPhysics schemas, and why USD beats SDF/URDF/MJCF for modern simulation pipelines. - [How to add physics properties to 3D models for simulation](https://rigyd.com/resources/how-to-add-physics-properties-to-3d-models-for-simulation/index.md): Adding physics, mass, friction, collision meshes, inertia, is the bottleneck in robotics simulation. Here's the manual workflow, the AI-automated alternative, and a step-by-step guide for both. ## Rankings - [Best Agentic Solutions for Creating Scenes and Objects for Simulation](https://rigyd.com/best/agentic-asset-and-scene-generation/index.md): Ranked: the four agentic platforms generating scenes and objects for robotics simulation in 2026 — Rigyd, Lightwheel, NVIDIA Edify, Moonlake. Strengths, limitations, and the right pick per use case. - [Best Simulators for Humanoid Robot Training](https://rigyd.com/best/humanoid-robot-simulators/index.md): Ranked: the simulators humanoid robotics teams actually use to train foundation-model policies in 2026 — Isaac Lab, MuJoCo MJX, Genesis, Gazebo. Throughput, accuracy, learning-stack integration, and the right pick per use case. - [Best 3D Asset Sources for Robotics Simulation](https://rigyd.com/best/3d-asset-sources-for-robotics-simulation/index.md): Ranked: where robotics teams actually get SimReady 3D assets in 2026 — agentic generation vs curated libraries vs research datasets. Coverage, calibration, format support, and the right source per workload. - [Best Benchmarks for Embodied AI and Humanoid Robotics](https://rigyd.com/best/embodied-ai-benchmarks-for-humanoid-robots/index.md): Ranked: the benchmarks humanoid and embodied-AI teams use to evaluate foundation models in 2026 — HumanoidBench, BEHAVIOR-1K, ManiSkill, Open X-Embodiment, RoboGen. Coverage, fidelity, and the right benchmark per use case. ## Asset categories - [SimReady Assets for Medical Equipment](https://rigyd.com/simready-assets/medical-equipment/index.md): Generate physics-enabled SimReady assets for medical equipment: surgical instruments, monitors, IV stands, patient beds, lab consumables. OpenUSD and MJCF, ready for Isaac Sim, MuJoCo, Genesis. - [SimReady Assets for Kitchen and Cooking](https://rigyd.com/simready-assets/kitchen/index.md): Generate physics-enabled SimReady assets for kitchen environments: cookware, utensils, fresh produce, appliances, dishware. OpenUSD and MJCF, ready for cooking-robot and home-manipulation simulators. - [SimReady Assets for Furniture](https://rigyd.com/simready-assets/furniture/index.md): Generate physics-enabled SimReady assets for furniture: chairs, tables, sofas, beds, cabinets, drawers. OpenUSD and MJCF for home robotics, assembly, and indoor navigation training. - [SimReady Assets for Home Appliances](https://rigyd.com/simready-assets/appliances/index.md): Generate physics-enabled SimReady assets for home appliances: washers, dryers, dishwashers, fridges, microwaves, ovens, vacuums. OpenUSD and MJCF for home robotics and assistive-robot training. ## Simulation platforms - [SimReady Assets for NVIDIA Isaac Sim](https://rigyd.com/for/isaac-sim/index.md): Generate physics-enabled OpenUSD assets optimized for NVIDIA Isaac Sim. Automatic mass, friction, and collision mesh estimation. Convert 3D models, images, or text and get SimReady USD in minutes. - [OpenUSD Assets for NVIDIA Omniverse](https://rigyd.com/for/omniverse/index.md): Create simulation-ready OpenUSD assets for NVIDIA Omniverse. Physics-enabled 3D models with mass, friction, collision meshes, and semantic labels. Compatible with Omniverse SimReady spec. - [Physically accurate Assets for Gazebo Simulation](https://rigyd.com/for/gazebo/index.md): Generate physically accurate 3D assets for Gazebo and ROS 2 robotics simulation. Automatic mass, friction, and collision geometry estimation from .glb, .fbx, .obj files. - [Simulation-Ready Models for MuJoCo](https://rigyd.com/for/mujoco/index.md): Create simulation-ready 3D models for MuJoCo with accurate mass, friction, and collision geometry. Convert .glb, .fbx, .obj to physics-enabled assets for robot learning. - [Digital Twin Assets for Unreal Engine](https://rigyd.com/for/unreal/index.md): Generate simulation-ready assets for Unreal Engine digital twins. Physics-enabled OpenUSD with mass, friction, and collision meshes. Import-ready for UE5 robotics simulation. - [Robotics Simulation Assets for Unity](https://rigyd.com/for/unity/index.md): Generate physically accurate 3D assets for Unity robotics simulation. Automatic mass, inertia, friction, and collision mesh estimation. Drop .fbx or OpenUSD into Unity Robotics Hub and ML-Agents. - [SimReady Assets for Genesis Simulator](https://rigyd.com/for/genesis/index.md): Generate physics-enabled assets for Genesis, the open-source GPU-native robotics simulator. Drop in 3D, images, or text descriptions and get MJCF or OpenUSD outputs ready for Genesis RL training. ## Robotics verticals - [Warehouse Simulation Assets](https://rigyd.com/verticals/warehouse-robots/index.md): Generate thousands of physically accurate warehouse objects for robotics simulation. Boxes, pallets, shelving, tools, all with mass, friction, and collision meshes for sim-to-real transfer. - [Robotic Arm Training Assets](https://rigyd.com/verticals/robotic-arm/index.md): Create physically accurate objects for robotic arm manipulation training. Automatic mass, friction, center of mass, and collision mesh generation for grasping, stacking, and assembly tasks. - [Humanoid Robot Simulation Assets](https://rigyd.com/verticals/humanoid-robots/index.md): Validated OpenUSD and MJCF SimReady assets for humanoid platforms — Atlas, Figure 02, 1X NEO, Apptronik Apollo, Unitree H1/G1, Tesla Optimus. Physical accuracy verified by test, collision geometry optimised for whole-body contact. Drop-in for NVIDIA Isaac Lab, GR00T, MuJoCo MJX and Genesis. - [SimReady Assets for Self-Driving](https://rigyd.com/verticals/autonomous-vehicles/index.md): Build AV simulation scenes with physically accurate 3D assets. Vehicles, pedestrians, road furniture, debris, all with mass, collision meshes, and semantic labels for self-driving perception and planning training. - [Drone Simulation Environments](https://rigyd.com/verticals/drones/index.md): Build drone simulation environments with physically accurate 3D assets. Obstacles, cargo, delivery targets, landing zones, with mass, collision meshes, and semantic labels for UAV autonomy training in Isaac Sim, AirSim, and Gazebo. - [Agritech Simulation Assets](https://rigyd.com/verticals/agricultural-robots/index.md): Generate physically accurate crops, implements, and farm environment assets for agricultural robot simulation. Mass, friction, and collision meshes for weeding, harvesting, and autonomous tractor training. - [SimReady Assets for Industrial Automation Robotics](https://rigyd.com/verticals/industrial-automation/index.md): Generate SimReady assets for industrial automation: production cells, assembly lines, fixturing, conveyors, parts bins. Rigyd converts 3D, images, or text into physics-enabled OpenUSD and MJCF. - [SimReady Assets for Robotics Research Labs](https://rigyd.com/verticals/research-labs/index.md): Generate SimReady assets for academic and industrial robotics research: benchmarks, ablation studies, sim-to-real transfer, foundation model training. Reproducible, citable, and works across Isaac Lab, MuJoCo, and Genesis. ## Use cases - [3D Assets for Better Sim-to-Real Transfer](https://rigyd.com/use-cases/sim-to-real/index.md): Close the sim-to-real gap with validated OpenUSD and MJCF SimReady assets. Rigyd derives mass, friction and collision meshes and verifies them by test, so policies trained in simulation transfer to hardware. - [Simulations Across Every Scenario and Edge Case](https://rigyd.com/use-cases/domain-randomization/index.md): Generate diverse, validated OpenUSD and MJCF assets for domain randomisation. Thousands of unique objects with physics verified by test, covering the variation a policy has to survive. - [3D to Digital Twin Pipeline](https://rigyd.com/use-cases/digital-twins/index.md): Turn 3D CAD and BIM models into physically accurate digital twins. Rigyd converts factory and warehouse models into SimReady OpenUSD with mass, friction, and collision meshes at thousands-of-objects scale. - [Physically Accurate Synthetic Data Generation](https://rigyd.com/use-cases/synthetic-data/index.md): Generate physically accurate synthetic training data for robotics. Rigyd builds the SimReady 3D asset layer, mass, friction, collision meshes, that Isaac Sim Replicator, Omniverse, and custom pipelines need for transferable datasets. - [RL Training Environments](https://rigyd.com/use-cases/reinforcement-learning/index.md): Build reinforcement-learning environments from validated OpenUSD and MJCF SimReady assets. Physics verified by test, collision geometry optimised so rollouts stay cheap, and an API and CLI to assemble the training world. - [SimReady Assets for Robot Policy Evaluation and Testing](https://rigyd.com/use-cases/evaluation-and-testing/index.md): Use Rigyd-generated SimReady assets for held-out policy evaluation, regression testing, safety scenarios, and pre-deployment validation. Reproducible, scalable, and matches the real environment. ## Posts - [Why we rebuilt Rigyd around an agent that reasons first](https://rigyd.com/posts/why-we-rebuilt-rigyd-around-an-agent-that-reasons-first/index.md): Three months, thousands of generations, and one lesson: the same object has to be built differently for pick-and-place than for dexterous manipulation. So we put reasoning ahead of the build. ## Optional - [Privacy Policy](https://rigyd.com/privacy-policy/index.md): How rigyd.com collects, uses and shares visitor information. - [User Agreement](https://rigyd.com/user-agreement/index.md): Terms governing access to and use of rigyd.com.