Kit-CAE User Guide#


Kit-CAE visualization examples showing vehicle aerodynamics, structural analysis, thermal simulation, and AI surrogate data

Kit-CAE is a reference implementation and developer toolkit for working with engineering simulation data using OpenUSD, NVIDIA CUDA-X libraries, and NVIDIA Omniverse. The Kit-CAE source code is available on GitHub. It demonstrates how to bring source of truth engineering data (for example, from simulation solvers, sensors, AI surrogate models) into a composable OpenUSD stage without converting it, then visualize and interact with that data using GPU-accelerated algorithms and RTX rendering.

The core idea: OpenUSD acts as the composition and access layer for scientific datasets while native files remain the source of truth. Kit-CAE implements this through a set of CAE schemas, data delegates, and visualization operators that you can use directly or adapt into your own workflows.

This guide walks you through the Kit-CAE application step by step. For deeper architectural detail (USD schemas, data delegate APIs, operator commands), see the reference documentation in the Kit-CAE repository.