Po3 MCP

🚀 Poe o3 MCP Server

A lightweight implementation of the Model Context Protocol (MCP) server that provides access to OpenAI's o3 model and other models through Poe's API. This server allows you to integrate Poe's AI capabilities into any MCP-compatible application.

🚀 Quick Start

This is a lightweight implementation of the Model Context Protocol (MCP) server. It offers access to OpenAI's o3 model and other models via Poe's API, enabling seamless integration of Poe's AI capabilities into MCP-compatible applications.

✨ Features

• A simple MCP server implemented with FastMCP.
• Direct integration with the Poe API to access the o3 model and other available models.
• Model selection in prompts using command-line style flags.
• Asynchronous request handling for improved processing efficiency.
• Comprehensive error handling and logging.
• Easy installation and configuration.

📦 Installation

Prerequisites

• Python 3.8+
• A Poe API key (obtainable from https://poe.com/api_key)

Steps

1. Clone this repository:

   git clone https://github.com/Anansitrading/po3_MCP.git
   cd po3_MCP
   

2. Create a virtual environment (recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install the required dependencies:

   pip install -r requirements.txt
   

4. Set up environment variables:

   cp sample.env .env
   

5. Edit the .env file and add your Poe API key:

   POE_API_KEY=your_poe_api_key_here
   

💻 Usage Examples

Running the MCP Server

Run the server as follows:

python poe_o3_mcp_server.py

The server will start and listen for MCP protocol messages on standard input/output.

Selecting a Model via Flags

You can select different models on Poe by adding flags to your prompts:

--Claude-3.5-Sonnet Tell me about quantum computing

This will route your query to the Claude-3.5-Sonnet model instead of the default o3 model.

The flag can be placed anywhere in the message:

• At the beginning: --GPT-4 What is the capital of France?
• In the middle: Tell me --Claude-3-Opus about the history of Rome
• At the end: What are the three laws of robotics? --Claude-3.5-Sonnet

The flag is automatically removed before the message is sent to the model.

If no flag is specified, the "o3" model is used by default.

Integrating with an MCP Client

This server provides two tools:

1. o3_query - Sends a query to the o3 model (or another model selected via a flag) and gets a response.
2. ping - A simple testing tool that returns "pong".

Here is an example of calling the tools using an MCP client:

from mcp.client import MCPClient

# Connect to the MCP server
client = MCPClient(server_command=["python", "path/to/poe_o3_mcp_server.py"])

# Call the o3_query tool, using the o3 model by default
response = client.call_tool("o3_query", {"message": "Tell me about quantum computing"})
print(response)

# Call the o3_query tool with a flag to select another model
response = client.call_tool("o3_query", {"message": "Hello, what is AI?"})
print(response)

# Call the ping tool to test the connection
response = client.call_tool("ping", {})
print(response)

📚 Documentation

Configuration

• POE_API_KEY: Required. The key for accessing the Poe API. Default is empty.
• MCP_SERVER_PORT: Optional. Sets the port the server listens on. Default is 5000.

Troubleshooting

1. API key not found:

   • Check if the POE_API_KEY environment variable is set correctly.

2. Connection issues:

   • Ensure the server is started and listening on the correct port.
   • Check the firewall settings to make sure the target port is open.

3. Dependency errors:

   • Ensure all required Python packages are installed. Refer to the requirements.txt file.

📄 License

This project is licensed under the MIT License. For more information, see the LICENSE file.

Acknowledgments

Thanks to the following technologies for their support and contributions:

• FastMCP
• Poe API
• The continuous support and collaboration of the open-source community