Multi AI Advisor

You run a Multi-Model Advisor MCP server that queries several local Ollama models in parallel, each with its own persona, and then combines their perspectives into a single, synthesized answer. This lets you explore diverse viewpoints on a topic and choose the most useful guidance from a council of AI advisors.

How to use

Ask Claude to leverage the Multi-Model Advisor to gather multiple model opinions and produce a synthesized response. You can request models with specific roles and then review the combined output that blends different viewpoints into a comprehensive answer.

How to install

Prerequisites you need before installation are Node.js 16.x or higher and Ollama installed and running. You should also have Claude for Desktop if you want the complete advisory experience.

Step 1: Install the MCP server for Claude Desktop automatically via the installer tool.

npx -y @smithery/cli install @YuChenSSR/multi-ai-advisor-mcp --client claude

Configuration and runtime setup

Configure the MCP server and environment variables to control model behavior and integration with Ollama.

Create a .env file in the project root with your desired settings. Example variables shown.

# Server configuration
SERVER_NAME=multi-model-advisor
SERVER_VERSION=1.0.0
DEBUG=true

# Ollama configuration
OLLAMA_API_URL=http://localhost:11434
DEFAULT_MODELS=gemma3:1b,llama3.2:1b,deepseek-r1:1.5b

# System prompts for each model
GEMMA_SYSTEM_PROMPT=You are a creative and innovative AI assistant. Think outside the box and offer novel perspectives.
LLAMA_SYSTEM_PROMPT=You are a supportive and empathetic AI assistant focused on human well-being. Provide considerate and balanced advice.
DEEPSEEK_SYSTEM_PROMPT=You are a logical and analytical AI assistant. Think step-by-step and explain your reasoning clearly.

Connect to Claude for Desktop

Add the MCP server as a local stdio process in Claude for Desktop so Claude can invoke it directly.

{
  "mcpServers": {
    "multi_model_advisor": {
      "command": "node",
      "args": ["/absolute/path/to/multi-ai-advisor-mcp/build/index.js"]
    }
  }
}

Replace "/absolute/path/to/" with the actual path to your project directory, then restart Claude for Desktop.

## Usage patterns

List available models to verify which Ollama models exist on your system and which are configured as defaults.

Show me which Ollama models are available on my system

## How it works

Two MCP tools are exposed: list-available-models and query-models. When you ask Claude to use the multi-model advisor, Claude invokes query-models, the server queries multiple Ollama models in parallel (each with its own persona), and Claude synthesizes a comprehensive answer from all responses.

## Troubleshooting

If Ollama cannot be reached, verify Ollama is running and the API URL is correct in your .env. You can test Ollama responsiveness by opening http://localhost:11434 in a browser.

If a model isn’t found, ensure you pulled it with Ollama and confirm the exact model name with ollama list. Use the list-available-models tool to see current availability.

If MCP tools do not appear in Claude, restart Claude after updating the configuration and confirm the absolute path in claude\_desktop\_config.json is correct. Check logs for errors.

If there is insufficient RAM for larger models, consider selecting smaller models or adjusting the configuration to reduce memory usage.

## Notes

The Multi-Model Advisor supports assigning different personas to each model to encourage diverse perspectives and richer synthesis of advice.

## Available tools

### list-available-models

Shows all Ollama models on your system and marks which ones are configured as defaults.

### query-models

Queries multiple models with a question and returns each model's perspective for synthesis.