MCP Auto Server

Project Overview and Purpose

The MCP Auto Server is a tool designed to help AI assistants select the most suitable Large Language Models (LLMs) based on cost efficiency and performance criteria. It leverages the Model Context Protocol (MCP) to provide a standardized interface for accessing and utilizing LLM selection logic.

Key Features:

• LLM Selection: Analyzes input prompts and recommends optimal LLMs based on factors like cost, token usage, and task suitability.
• MCP Compliance: Implements the Model Context Protocol for seamless integration with AI assistants and other MCP-compatible tools.
• Configuration: Allows customization of server settings, including rate limiting and caching options, via the mcp_settings.json file.
• Cline Integration: Designed for use with the Cline AI assistant, providing a streamlined user experience.

Installation Instructions

1. Prerequisites:

   • Node.js (version 18 or higher)
   • npm (Node Package Manager)

2. Installation Steps:

   • Clone the repository:

     git clone <repository_url>
     cd mcp-auto
     

   • Install dependencies:

     npm install
     

   • Configure the server:
     • Review and modify the mcp_settings.json file to adjust server settings as needed.

3. Verification:

   • Start the server:

     npm start
     

   • Verify that the server is running without errors.

Configuration Details

The mcp_settings.json file contains detailed configuration options for the MCP Auto Server.

Key Configuration Options:

• Rate Limiting: Configure the maximum number of requests allowed per time window to prevent abuse and ensure server stability.
• Caching: Enable caching to store frequently accessed LLM selection results, improving performance and reducing costs.
• Other Server Settings: Customize other server parameters, such as port number, host address, and logging levels.

Example Configuration:

{
  "port": 3000,
  "host": "localhost",
  "rateLimit": {
    "windowMs": 60000, // 1 minute
    "max": 100 // 100 requests per minute
  },
  "cache": {
    "enabled": true,
    "ttl": 3600 // 1 hour
  }
}

Usage Examples

The MCP Auto Server is designed to be used with the Cline AI assistant. Users do not directly interact with the server; instead, they use Cline, which connects to the server automatically when needed.

Example Usage:

1. Invoke Cline and provide a prompt:

   Cline: "Which LLM is best for writing a short story about a robot learning to paint?"
   

2. Cline will connect to the MCP Auto Server, analyze the prompt, and recommend the most suitable LLM.

3. Cline will then use the recommended LLM to generate the short story.

API Documentation

The MCP Auto Server implements the following MCP tools:

analyze_prompt

Analyzes an input prompt and recommends the most suitable LLM models based on cost efficiency and performance criteria.

Input:

{
  "prompt": "string" // The input prompt to analyze
}

Output:

{
  "inputPrompt": "string",
  "estimatedTokenCount": "number",
  "detectedTasks": "array",
  "recommendations": {
    "topModels": "array",
    "optimizationTips": "array",
    "optimizedPromptExample": "string"
  }
}

Error Handling:

The server returns standard MCP error responses for invalid requests or internal server errors.

Troubleshooting Guide

Common Issues:

• Server not starting:
  • Verify that Node.js and npm are installed correctly.
  • Check the console for error messages.
  • Ensure that the port specified in mcp_settings.json is not already in use.
• Cline not connecting to the server:
  • Verify that the server is running.
  • Check the Cline configuration to ensure that it is pointing to the correct server address.
  • Inspect the Cline logs for connection errors.

Logging and Debugging:

The MCP Auto Server uses a custom logging solution. Check the server logs for detailed information about server activity and errors.

Support Resources:

• Contact the development team for assistance.