---
title: "NVIDIA Integration"
source: https://docs.autohand.ai/integrations/nvidia
---

# NVIDIA

Connect Autohand to NVIDIA-hosted models through NVIDIA's OpenAI-compatible inference endpoint.

## Overview

NVIDIA is available as the `nvidia` provider. Use it when your team wants NVIDIA-hosted model access, GPU-backed inference infrastructure, or a provider catalog that includes NVIDIA, Mistral, DeepSeek, and related model families.

## Setup

``` bash
export NVIDIA_API_KEY="your-nvidia-api-key"

autohand --provider nvidia --model deepseek-ai/deepseek-v4-pro
```

## CLI configuration

``` json
{
  "provider": "nvidia",
  "nvidia": {
    "apiKey": "${NVIDIA_API_KEY}",
    "baseUrl": "https://integrate.api.nvidia.com/v1",
    "model": "deepseek-ai/deepseek-v4-pro"
  }
}
```

| Option | Description | Default |
|---|---|---|
| apiKey | Your NVIDIA API key | - |
| baseUrl | NVIDIA inference endpoint | https://integrate.api.nvidia.com/v1 |
| model | NVIDIA-hosted model ID | deepseek-ai/deepseek-v4-pro |

## Models

Select NVIDIA-hosted models with `/model` or configure a custom model ID from your NVIDIA account.

| Example model | Best for |
|---|---|
| deepseek-ai/deepseek-v4-pro | Advanced coding and reasoning |
| mistralai/mixtral-8x7b-instruct-v0.1 | General-purpose instruction following |
| nvidia/usdcode | NVIDIA-specific coding and content workflows |