Anvil Physics
Rigid bodies, soft bodies, cloth and fluids — differentiable.
Physics engine with a differentiable mode for policy gradient training, deterministic replay, and articulation solvers tuned for contact-rich manipulation rather than for looking plausible in a game.
Specifications
| Solvers | Rigid, soft, cloth, SPH/FLIP fluid |
|---|---|
| Differentiable | Yes, gradients through contact |
| Determinism | Bitwise replay on same build |
| Scale | 40,000 parallel environments |
| Family | Simulation & Digital Twins |
| Software stack | FLUX — source-compatible across the whole catalog |
Disclosure
Anvil Physics is invented. The numbers above were written to be internally consistent and plausible, not measured. Treat every figure on this site as placeholder content for layout, search and API testing.
Related
Often deployed together
AXION Loom
One scene graph, every DCC tool.
- Format
- OpenUSD native
- Live editing
- Multi-user, multi-app stage
Loom Kit SDK
Build your own tool on the stage.
- Languages
- Python, C++
- Modes
- Desktop app, headless service
Proving Ground
Train and test robots where breaking them is free.
- Sensors
- Camera, depth, lidar, IMU, F/T — calibrated
- Throughput
- 40,000 envs on one M400
Mint Synthetic Data
Labeled data for the cases your fleet never sees.
- Outputs
- RGB, depth, segmentation, boxes, lidar
- Labels
- Perfect by construction