# Rigyd > Rigyd is AI-native 3D infrastructure for robotics learning and evaluation. It converts raw 3D models, images, and text into physics-enabled, simulation-ready OpenUSD and MJCF assets. Rigyd is a member of the NVIDIA Inception Program for AI startups. ## What Rigyd Does Rigyd solves the 3D data bottleneck in robotics simulation. Most 3D assets lack the physical properties (mass, friction, collision geometry) that simulators need. Rigyd uses AI to automatically: - Estimate physics properties: mass, friction, restitution, center of mass - Generate collision meshes via convex decomposition - Add semantic labels and material identification - Output validated OpenUSD with USDPhysics schemas (PhysicsRigidBodyAPI, PhysicsMassAPI, PhysicsCollisionAPI) - Export the same asset to MuJoCo MJCF (.xml) with inertials, geoms, friction, and solref contact parameters ## Input Formats - .glb (glTF Binary) - .fbx (Autodesk FBX) - .obj (Wavefront OBJ) - Images - Text descriptions ## Output Formats - OpenUSD (.usda / .usdc) with full USDPhysics schemas - MuJoCo MJCF (.xml) with inertials, collision geoms, friction, and contact parameters ## Compatible Simulators - NVIDIA Isaac Sim (via OpenUSD) - NVIDIA Omniverse (asset-validated) - MuJoCo (Google DeepMind) - Gazebo / ROS 2 (Open Robotics) - Unreal Engine (Epic Games) - Unity ## Key Statistics - Only ~1,000 SimReady assets exist today in NVIDIA's library; typical warehouses have 50,000+ SKUs. - 40% better real-world performance when training combines synthetic + real data with physically accurate assets (Source: NVIDIA GR00T N1 Benchmark, 2025). - 97% cost reduction in scene preparation vs manual workflows. - ~$370K saved per 1,000-object simulation project. ## Three Tracks 1. **3D to SimReady** (Available now): Upload existing 3D models. Rigyd optimizes geometry, adds physics properties, collision meshes, semantic labels, and packages as OpenUSD plus MJCF. 2. **Text / Image to SimReady** (Available now): Describe what you need or upload images. AI generates 3D geometry with the required topology, textures, collision meshes, semantic labels, and physics properties. 3. **Worlds to SimReady** (Coming soon): Describe the environments your robots need. AI orchestrates world models to generate diverse, physically accurate 3D scenes with collidable meshes. ## How It Works 1. **Upload**: drop any 3D file (.glb, .fbx, .obj) or describe what you need. 2. **AI estimates physics**: generates collision meshes, identifies materials, estimates mass, friction, center of mass, restitution. 3. **Download SimReady USD and MJCF**: validated OpenUSD with USDPhysics schemas plus a MJCF export, ready for Isaac Sim, MuJoCo, Gazebo, Unreal Engine, and Unity. ## Resources ### Compare hub - Compare index: https://rigyd.com/compare - Rigyd vs NVIDIA SimReady library: https://rigyd.com/compare/rigyd-vs-nvidia-simready-library - Isaac Sim vs MuJoCo vs Gazebo: https://rigyd.com/compare/isaac-sim-vs-mujoco-vs-gazebo - OpenUSD vs SDF vs URDF vs MJCF: https://rigyd.com/compare/openusd-vs-sdf-urdf-vs-mjcf - Manual vs AI physics annotation: https://rigyd.com/compare/manual-vs-ai-physics-annotation ### Resources hub - Resources index: https://rigyd.com/resources ### Engine pages (/for/{engine}) - https://rigyd.com/for/isaac-sim - https://rigyd.com/for/omniverse - https://rigyd.com/for/mujoco - https://rigyd.com/for/gazebo - https://rigyd.com/for/unreal - https://rigyd.com/for/unity ### Vertical pages (/verticals/{type}) - https://rigyd.com/verticals/warehouse-robots - https://rigyd.com/verticals/humanoid-robots - https://rigyd.com/verticals/autonomous-vehicles - https://rigyd.com/verticals/drones - https://rigyd.com/verticals/robotic-arm - https://rigyd.com/verticals/agricultural-robots ### Use-case pages (/use-cases/{case}) - https://rigyd.com/use-cases/sim-to-real - https://rigyd.com/use-cases/domain-randomization - https://rigyd.com/use-cases/synthetic-data - https://rigyd.com/use-cases/reinforcement-learning - https://rigyd.com/use-cases/digital-twins ## Links - Website: https://rigyd.com - App: https://app.rigyd.com - X: https://x.com/rigyd3d - YouTube: https://www.youtube.com/@rigyd3d - Privacy Policy: https://rigyd.com/privacy-policy - User Agreement: https://rigyd.com/user-agreement ## Contact - Request API Access: use the Request API Access form linked in the site navigation at https://rigyd.com - Email: support@rigyd.com