Multi Model Advisor

🚀 Multi-Model Advisor

The Multi-Model Advisor is a server based on the MCP protocol. It aims to provide more comprehensive solutions by querying multiple Ollama models and combining their responses. This server allows you to specify different roles and perspectives, enabling the generation of richer and more diverse answers.

🚀 Quick Start

The Multi-Model Advisor is a powerful tool that combines multiple Ollama models to offer comprehensive solutions. Here's a quick guide to get you started.

✨ Features

• Multi-Model Query: Supports simultaneous invocation of multiple pre-trained language models.
• Role Assignment: Each model can be assigned different roles or tasks.
• Available Model List: Lists all installed Ollama models on the local computer.

📦 Installation

Prerequisites

1. Ensure Node.js (version 16.x or higher) is installed.
2. Install and run the Ollama service:

   ollama serve
   

3. Pull the required models:

   ollama pull gemma3:1b llama3.2:1b deepseek-r1:5b
   

Installation Methods

Method 1: Install via Smithery CLI

# Install the smithery-cli tool
npm install -g smithery-cli

# Build the project using smithery
smithery build git@github.com:YuChenSSR/multi-ai-advisor-mcp.git

Method 2: Manual Installation

1. Clone the project repository:

   git clone https://github.com/YuChenSSR/multi-ai-advisor-mcp.git
   cd multi-ai-advisor-mcp
   

2. Install dependencies:

   npm install
   

3. Start the server:

   node index.js
   

📚 Documentation

Configuration Guide

Create a .env file and add the following content:

# Service configuration
PORT=3000
API_BASE_URL=http://localhost:11434

# System prompt examples
GEMMA3_PERSONA=Gemma is a professional career advisor.
LLAMA3_PERSONA=Llama is an experienced technical expert.
DEEPSEEK_R1_PERSONA=DeepSeek-R1 is a strategic planner.

Usage Instructions

Basic Usage

Enter the following query in Claude:

what are the most important skills for success in today's job market, 
you can use gemma3:1b, llama3.2:1b, deepseek-r1:5b to help you

Function Example

You can use gemma3:1b, llama3.2:1b, and deepseek-r1:5b to help you analyze what the most crucial skills are in today's job market.

Working Principle

1. Tool Exposure:
   • list-available-models: Displays all local Ollama models.
   • query-models: Sends questions to multiple models.
2. Query Process:
   • When you pose a question in Claude, Claude calls the query-models tool.
   • The server forwards the question to the specified models.
   • Each model provides an answer based on its role.
   • The final answer is generated by Claude integrating the outputs of all models.

Troubleshooting

Ollama Connection Issues

• Ensure the Ollama service is started: ollama serve
• Verify that OLLMAR_API_URL is correct.
• Test access to http://localhost:11434 to confirm that Ollama is running properly.

Model Unavailability

• Use ollama list to check the model status.
• Ensure the required models are pulled: ollama pull <model-name>

Claude Not Displaying MCP Tools

• Restart the Claude client.
• Confirm that the configuration file path is correct.

📄 License

This project is licensed under the MIT License. Please refer to the LICENSE file for specific terms.

👥 Contribution Guide

Community contributions are welcome! Please refer to CONTRIBUTING.md to learn how to participate in the development.