n8n-nodes-nvidia 🟢

A custom n8n node for the NVIDIA NIM API — supports chat completions, streaming, thinking mode and all NIM-compatible models (Kimi K2.5, Llama, Mistral, etc.)

Features

• ✅ OpenAI-compatible Chat Completions
• ✅ Streaming support (SSE / text/event-stream)
• ✅ Thinking mode (chat_template_kwargs)
• ✅ All NVIDIA NIM models selectable
• ✅ Full credential management in n8n
• ✅ Configurable: temperature, top_p, max_tokens

Installation

Option A — Custom Nodes Folder (recommended for self-hosted n8n)

# 1. Find your n8n custom nodes folder (usually):
#    ~/.n8n/custom  OR  /home/<user>/.n8n/nodes

cd ~/.n8n/nodes   # create if it doesn't exist: mkdir -p ~/.n8n/nodes

# 2. Clone / copy this package
git clone https://github.com/julius/n8n-nodes-nvidia.git
cd n8n-nodes-nvidia

# 3. Install dependencies & build
npm install
npm run build

# 4. Tell n8n where to find it (in your .env or n8n config):
#    N8N_CUSTOM_EXTENSIONS=/home/<user>/.n8n/nodes/n8n-nodes-nvidia/dist

# 5. Restart n8n

Option B — npm link (for development)

cd n8n-nodes-nvidia
npm install
npm run build
npm link

# In your n8n folder:
npm link n8n-nodes-nvidia

Option C — Publish to npm, then install via n8n UI

npm publish --access public

# In n8n UI: Settings → Community Nodes → Install → "n8n-nodes-nvidia"

Configuration

Credential Setup

1. In n8n: Credentials → New → NVIDIA NIM API
2. Paste your API key from https://integrate.api.nvidia.com/
3. Hit "Test" to verify

Node Parameters

| Parameter | Description | Default | |-----------|-------------|---------| | Model | NIM model identifier | moonshotai/kimi-k2.5 | | Messages | Chat message list (role + content) | user message | | Max Tokens | Max tokens to generate | 16384 | | Temperature | Creativity (0–2) | 1.0 | | Top P | Nucleus sampling (0–1) | 1.0 | | Enable Streaming | Use SSE streaming | false | | Enable Thinking | Adds chat_template_kwargs: {thinking: true} | false | | Custom chat_template_kwargs | Raw JSON override | — | | System Prompt Shortcut | Quick system message | — | | Return Raw Response | Include full API response | false |

Popular Models

| Model | Description | |-------|-------------| | moonshotai/kimi-k2.5 | Kimi K2.5 with thinking support | | meta/llama-3.1-70b-instruct | Llama 3.1 70B | | meta/llama-3.3-70b-instruct | Llama 3.3 70B | | mistralai/mistral-large-2-instruct | Mistral Large 2 | | nvidia/llama-3.1-nemotron-70b-instruct | NVIDIA Nemotron | | google/gemma-3-27b-it | Gemma 3 27B |

Browse all: https://build.nvidia.com/explore/discover

Example Workflow

{
  "nodes": [
    {
      "name": "NVIDIA NIM LLM",
      "type": "nvidiaLlm",
      "parameters": {
        "model": "moonshotai/kimi-k2.5",
        "messages": {
          "messagesValues": [
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": "Explain quantum entanglement simply."}
          ]
        },
        "maxTokens": 4096,
        "temperature": 0.7,
        "stream": false,
        "advancedOptions": {
          "thinking": true
        }
      }
    }
  ]
}

Output

The node outputs:

{
  "text": "The model's response text here...",
  "model": "moonshotai/kimi-k2.5",
  "stream": false,
  "usage": {
    "prompt_tokens": 42,
    "completion_tokens": 512,
    "total_tokens": 554
  }
}

Development

npm run dev      # watch mode
npm run lint     # lint check
npm run lint:fix # auto-fix lint
npm run build    # production build

MIT License — Built by Julius @ funnyareas.com