jedarden/gpt-4o-search-mcp

Project Overview

A Model Context Protocol (MCP) server which makes OpenAI's gpt-4o-search-preview model accessible over MCP.

• : Main application entry point.
• : Lists Python dependencies required to run the application.
• : Instructions for building and running the application in a Docker container.
• : Example environment variables file. Copy this to .env and update values as needed.

Deployment Instructions

1. Environment Variables

Before running the application, set up your environment variables:

1. Copy .env.example to .env:

   cp .env.example .env
   

2. Edit .env and update the values as needed for your environment.

2. Deploying with Docker

1. Build the Docker image:

   docker build -t my-python-app -f dockerfile .
   

2. Run the container:

   docker run --env-file .env -p 8000:8000 my-python-app
   

3. Deploying with Python (virtualenv)

1. Create and activate a virtual environment:

   python3 -m venv venv
   source venv/bin/activate
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Set environment variables (see .env.example).
4. Run the application:

   python app/app.py
   

Example: Using roo code to connect to MCP

Below is an example configuration block for the gpt-4o-search MCP service:

"gpt-4o-search": {
  "url": "http://link-to-where-service-is-hosted:8000/sse",
  "transport": "http",
  "alwaysAllow": [
    "search"
  ],
  "timeout": 300
}

Python Example: Performing a "search" Operation

The following Python code demonstrates how to use the above configuration to connect to the MCP service and perform a "search" operation using roo code principles. This example uses the requests library to send a search request to the MCP endpoint.

from mcp import MCPClient

# Initialize the MCP client for the gpt-4o-search server
client = MCPClient("http://link-to-where-service-is-hosted:8000/sse")

# Perform a "search" operation
result = client.tool("search", {"query": "What is Model Context Protocol?"})

print("Search result:", result)
### Explanation

- **MCPClient**: The official `mcp` Python library provides the `MCPClient` class to connect to an MCP server.
- **client = MCPClient(...)**: Initializes the client with the URL of the gpt-4o-search MCP server.
- **client.tool("search", {...})**: Performs the "search" operation by specifying the tool name and parameters as a dictionary.
- **Result**: The result of the search operation is printed.

---
# Notes

- Only perform the work outlined above and do not deviate from these instructions.
- For further details, refer to the individual files and comments within the codebase.