n8n-nodes-rooyai-message

A production-ready Rooyai Message / Chat Model node for n8n, providing first-class LLM provider integration equivalent to OpenAI, Gemini, or DeepSeek.

🎯 Overview

This custom n8n community node enables you to use Rooyai's LLM API as a message/chat model provider in your n8n workflows. It appears under AI → Language Models → Rooyai Message Model and works seamlessly with:

• ✅ AI Agent
• ✅ Better AI Agent
• ✅ Basic LLM Chain
• ✅ Tools
• ✅ Memory

📦 Installation

Option 1: Install in n8n Custom Directory (Recommended for Testing)

# Create custom nodes directory if it doesn't exist
mkdir -p ~/.n8n/custom

# Copy the entire dist folder to the custom directory
cp -r ./dist ~/.n8n/custom/n8n-nodes-rooyai-message

# Restart n8n
n8n restart

Option 2: Install via npm (Production)

# In your n8n installation directory
npm install n8n-nodes-rooyai-message

# Restart n8n
n8n restart

Option 3: Development Link

# In this project directory
npm run build
npm link

# In your n8n directory
npm link n8n-nodes-rooyai-message
n8n restart

🔑 Credentials Setup

1. In n8n, navigate to Credentials → Create New Credential
2. Search for "Rooyai API"
3. Configure the following fields:

| Field | Type | Required | Description | |-------|------|----------|-------------| | API Key | Password | ✅ Yes | Your Rooyai API authentication key | | Base URL | String | ✅ Yes | API endpoint (default: https://rooyai.com/api/v1/chat) | | Optional Headers | JSON String | ❌ No | Additional headers in JSON format: {"X-Custom": "value"} |

4. Click Save to store your credentials

🚀 Usage

Basic Chat Completion

1. Add Rooyai Message Model node to your workflow
2. Select your Rooyai API credentials
3. Configure the node:
   • Model: Select from dropdown (15 models available)
   • Messages: Add user/system/assistant messages
   • Temperature: 0.7 (0-2 range)
   • Max Tokens: 1024 (optional)

Example Workflow

Start Node → Rooyai Message Model → Output Node

Configuration:

• Model: LLaMa 3.3 70B (from dropdown)
• Messages:
  • Role: system, Content: You are a helpful assistant
  • Role: user, Content: Explain quantum computing in simple terms
• Temperature: 0.7

With AI Agent

Manual Chat Trigger → AI Agent → Rooyai Message Model

The Rooyai Message Model node integrates directly as a language model provider in AI Agent workflows.

With Basic LLM Chain

Start → Basic LLM Chain → Rooyai Message Model → Output

Configure the chain with your prompt template, and it will automatically use Rooyai for text generation.

⚙️ Configuration Options

Model Selection

Select from 15 available Rooyai models via dropdown:

| Model | Description | Best For | |-------|-------------|----------| | LLaMa 3.3 70B | Meta's flagship model with 70B parameters | Complex reasoning, detailed analysis | | DeepSeek R1 | Reasoning-optimized model | Logical tasks, problem-solving | | DeepSeek v3.1 Nex | Latest DeepSeek with enhancements | General purpose, advanced tasks | | Qwen3 Coder | Code generation specialist | Programming, technical documentation | | GPT OSS 120B | Large open-source GPT | Complex tasks, high accuracy | | GPT OSS 20B | Efficient open-source GPT | Fast responses, good balance | | TNG R1T Chimera | TNG reasoning architecture | Analytical tasks | | TNG DeepSeek Chimera | Hybrid TNG-DeepSeek model | Multi-domain tasks | | Kimi K2 | Moonshot AI's multilingual model | Chinese language, translations | | GLM 4.5 Air | Lightweight ChatGLM | Fast interactions, efficiency | | Devstral | Developer-focused model | Coding, debugging, tech docs | | Mimo v2 Flash | High-speed model | Quick responses, real-time chat | | Gemma 3 27B | Google Gemma large variant | General purpose, quality | | Gemma 3 12B | Google Gemma balanced | Good performance/speed ratio | | Gemma 3 4B | Google Gemma compact | Fastest responses, simple tasks |

Message Roles

• system: Defines AI behavior and context
• user: Human input/questions
• assistant: AI responses (for conversation history)

Advanced Options

| Option | Type | Range | Description | |--------|------|-------|-------------| | Temperature | Number | 0-2 | Controls randomness (0=deterministic, 2=very creative) | | Max Tokens | Number | 1-32768 | Maximum response length | | Frequency Penalty | Number | -2 to 2 | Reduces word repetition | | Presence Penalty | Number | -2 to 2 | Encourages new topics | | Top P | Number | 0-1 | Nucleus sampling (alternative to temperature) |

Simplify Output

• Enabled (default): Returns only the assistant's message content as a clean string
• Disabled: Returns full API response including usage metadata (cost_usd)

🔧 API Integration Details

Request Format

The node sends POST requests to your configured Base URL with:

{
  "model": "gemini-2.0-flash",
  "messages": [
    { "role": "system", "content": "You are helpful" },
    { "role": "user", "content": "Hello!" }
  ],
  "temperature": 0.7,
  "max_tokens": 1024
}

Headers:

Authorization: Bearer {YOUR_API_KEY}
Content-Type: application/json
{...optional custom headers}

Response Parsing

Rooyai returns responses in this format:

{
  "choices": [
    {
      "message": {
        "content": "Hello! How can I assist you today?"
      }
    }
  ],
  "usage": {
    "cost_usd": 0.000123
  }
}

The node automatically extracts choices[0].message.content for the final output.

📁 Project Structure

n8n-nodes-rooyai-message/
├── credentials/
│   └── RooyaiApi.credentials.ts    # API credentials definition
├── nodes/
│   └── RooyaiMessage/
│       ├── RooyaiMessage.node.ts   # Main node implementation
│       ├── ChatDescription.ts       # Message/chat operations
│       ├── GenericFunctions.ts      # Error handling & utilities
│       ├── RooyaiMessage.node.json  # Node metadata
│       └── rooyai.svg              # Node icon
├── dist/                           # Compiled JavaScript output
├── package.json                    # Package metadata & dependencies
├── tsconfig.json                   # TypeScript configuration
├── gulpfile.js                     # Build tasks (icon copying)
└── README.md                       # This file

🛠️ Development

Prerequisites

• Node.js 18+
• npm 8+
• TypeScript 5.3+

Build from Source

# Install dependencies
npm install

# Build the project (compiles TypeScript + copies icons)
npm run build

# Watch mode for development
npm run dev

Modifying the API Integration

⚙️ Change Base URL:
Edit credentials/RooyaiApi.credentials.ts, line 20:

default: 'https://your-new-endpoint.com/api/v1/chat'

⚙️ Modify Response Parsing:
Edit nodes/RooyaiMessage/ChatDescription.ts, lines 140-160 (postReceive function):

// Update to match your API's response structure
const assistantText = item.json?.choices?.[0]?.message?.content || '';

⚙️ Add Custom Headers:
Users can add custom headers via the "Optional Headers" credential field without code changes.

✅ Verification

After installation, verify the node:

1. Node Appears: Search for "Rooyai" in n8n's "Add Node" menu
2. Credentials Work: Create credential and test with valid API key
3. Chat Works: Send a test message and receive response
4. No Errors: Check n8n logs for any error messages

Expected behavior:

• Node is categorized under AI or Language Models
• Requests sent to configured Base URL
• Responses parsed correctly as strings
• Compatible with AI Agent and LLM Chain nodes

🐛 Troubleshooting

Node doesn't appear in n8n

• Ensure dist/ folder is copied to ~/.n8n/custom/
• Restart n8n: n8n restart or service n8n restart
• Check n8n logs: ~/.n8n/logs/n8n.log

"Cannot find credentials" error

• Create "Rooyai API" credentials in n8n UI first
• Ensure API key is valid and not expired

API request fails

• Verify Base URL is correct: https://rooyai.com/api/v1/chat
• Check API key has proper permissions
• Review error message in n8n execution view

Response parsing error

• Enable "Simplify Output: false" to see raw API response
• Verify Rooyai API returns choices[0].message.content

📝 License

MIT

👥 Author

Rooyai
Website: https://rooyai.com
Support: [email protected]

🔗 Links

• Rooyai API Documentation
• n8n Community Nodes
• n8n AI Workflows

Built with ❤️ for the n8n community