MCP Server Openai

🚀 OpenAI MCP Server

This project enables direct queries to OpenAI models from Claude via the MCP protocol. This branch adds support for the o3-mini and gpt-4o-mini models and improves message handling.

🚀 Quick Start

📦 Installation

1. Clone the repository:

   git clone https://github.com/thadius83/mcp-server-openai.git
   cd mcp-server-openai
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Run the server:

   python -m src.openai_mcp_server --openai-api-key your_api_key_here
   

Configuration

Add the following content to clint_mcp_settings.json:

{
  "servers": {
    "github.com/thadius83/mcp-server-openai": {
      "type": "tcp",
      "host": "localhost",
      "port": 5000
    }
  },
  "tools": {
    "ask-openai": {
      "server": "github.com/thadius83/mcp-server-openai",
      "tool": "ask-openai"
    }
  }
}

✨ Features

1. Supported models:

   • o3-mini (default): Suitable for quick and concise responses.
   • gpt-4o-mini: Ideal for more comprehensive and detailed explanations.

2. Improved message format:

   • Responses are returned in a structured JSON format, including type and content.

💻 Usage Examples

Use the following command in Claude:

<use_mcp_tool>
<server_name>github.com/thadius83/mcp-server-openai</server_name>
<tool_name>ask-openai</tool_name>
<arguments>
{
  "query": "What are the key features of Python's asyncio library?",
  "model": "o3-mini" // Optional, default is o3-mini
}
</arguments>
</use_mcp_tool>

Model Comparison

1. o3-mini (default):

   • Best for: Quick and concise responses.
   • Style: Direct and efficient.

2. gpt-4o-mini:

   • Best for: More comprehensive explanations.
   • Style: Detailed and thorough.

Response Format

The standard format for tool responses:

{
  "content": [
    {
      "type": "text",
      "text": "The model's response..."
    }
  ]
}

Troubleshooting

1. Server not found:

   • Ensure the PYTHONPATH in the configuration file is correct.
   • Check if Python and pip are installed correctly.

2. Authentication error:

   • Confirm that the OpenAI API key is valid.

3. Model error:

   • Ensure you are using a supported model (o3-mini or gpt-4o-mini).

Development Instructions

# Install development dependencies
pip install -e ".[dev]"

# Run tests
pytest -v test_openai.py -s

📄 Changelog

• Added support for the o3-mini and gpt-4o-mini models.
• Improved the message handling mechanism.