Crashcat – Lightweight 3D physics for JavaScript
Pure JS physics with CCD and shapecasting, skips the 3MB WASM tax.
Compact implementation of spatial algebra, and Featherstone's articulated-body algorithm (ABA), composite rigid body algorithm (CRBA), and the recursive Newton-Euler algorithm (RNEA) with end-to-end automatic differentiability (AD). Depends on STL, GLM and autodiff libraries.
C++ rigid body simulation with end-to-end automatic differentiation for ML.
Robotics engineers, graphics programmers, ML researchers
MuJoCo · Brax · Bullet Physics
Pure JS physics with CCD and shapecasting, skips the 3MB WASM tax.
LLM-powered decompilation with differential testing beats guesswork, but verification burden remains.
Directory-level differential sync that rsync and xdelta never solved for cloud storage.
Another neurosymbolic DB claiming to fix LLM hallucinations with zero adoption yet.
Solid learning project, but dual numbers for auto-diff already exist in better libraries.
C++ + OpenGL ES compiled to WASM—browser physics sim with predictive orbits.