# NVIDIA USDCode

> Use this tool when you need to generate expert-level code and scripting guidance for NVIDIA Omniverse, Isaac Sim, and USD-based 3D content pipelines. It provides direct access to NVIDIA's specialized USDCode Llama 3.1 70B model, solving problems related to scripting, API usage, and technical implementation details. The tool accepts configurable parameters such as temperature and token limits, and outputs tailored guidance and code snippets through a Python API interface.

Canonical page: https://skillsregistry.net/skills/nvidia-usdcode  
JSON: https://api.skillsregistry.net/v1/skills/nvidia-usdcode

## Description

NVIDIA USDCode MCP server by ishandotsh that provides AI assistants with direct access to NVIDIA's specialized USDCode Llama 3.1 70B model through the NVIDIA API, designed specifically for Isaac Sim scripting, USD (Universal Scene Description) workflows, and Python API assistance. Built with TypeScript using the OpenAI SDK for API compatibility, the implementation offers configurable parameters including temperature, top-p sampling, token limits, expert type selection (auto, knowledge, code, or helperfunction), and streaming support, with dynamic module loading to handle MCP SDK imports. Serves developers working with NVIDIA Omniverse, Isaac Sim robotics simulations, and USD-based 3D content pipelines who need expert-level guidance on scripting, API usage, and technical implementation details from a model trained specifically on NVIDIA's ecosystem and documentation.

## Trust

- **Trust score (0–1):** 0.99
- **Verification tier:** verified
- **Last scanned:** 2026-09-28

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** devops-ci
- **Updated:** 2026-09-28

## Source

- **Source listing:** [PulseMCP](https://www.pulsemcp.com/servers/nvidia-usdcode)
- **Repository:** <https://github.com/ishandotsh/nvidia-usdcode-mcp-server>

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "nvidia-usdcode"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/nvidia-usdcode` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/nvidia-usdcode/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt