USD Search NIM

USD Search NIM is an NVIDIA NIM inference microservice enabling developers, creators, and workflow specialists to take advantage of AI-powered search for OpenUSD data, 3D models, images, and assets using text or image-based inputs.

By combining a collection of cloud-native microservices, USD Search NIM helps developers and the users of their applications to save time navigating unstructured, untagged 3D data. It is capable of searching and indexing 3D asset databases, searching in-scene, and performing spatial searches, without requiring any manual tagging of assets.

To try USD Search NIM, check out the NVIDIA API Catalog, where it is currently available with a sample API endpoint.

Tip

To learn how to use USD Search NIM in your workflows, check out the Generative AI for Digital Twins Guide, where we share examples of using NVIDIA USD NIM microservices to build digital twins and synthetic data generation pipelines for training physical AI.

ILM leveraged USD Search NIM (previously DeepSearch) to sift through their massive database of 3D scenes and sky images. Watch a video about their process here