AI-Assistant-OpenAI-MCP-Setup

An AI assistant setup with OPENAI & MCP Server (Model Context Protocol) tool support and a Gradio chat UI. Built with FastAPI, FastMCP, and Gradio.This setup is fully customizable—follow the instructions to adapt it to your own environment.

User ↔ Gradio Chat UI ↔ OpenAI API ↔ MCP Server (FastAPI + FastMCP)
         /chat                            /mcp

• Gradio serves the chat interface at /chat
• OpenAI Responses API handles conversation and decides when to call tools
• MCP Server exposes tools via the /mcp endpoint using FastMCP
• Tool definitions are auto-discovered from MCP and passed to OpenAI as function tools

git clone https://github.com/febeeh/AI-Assistant-OpenAI-MCP-Setup
cd AI-Assistant-OpenAI-MCP-Setup

python -m venv venv
source venv/bin/activate        # Linux/Mac
venv\Scripts\activate           # Windows

pip install -r requirements.txt

Create a .env file in the project root:

OPENAI_API_KEY=sk-your-key-here
MODEL=gpt-4o-mini
HOST=localhost
PORT=8000

python main.py

• Chat UI: http://localhost:8000/chat
• MCP endpoint: http://localhost:8000/mcp

Edit assistant.md — update the system prompt:

Add new routes in api/mcp/router.py. Each FastAPI endpoint under the router automatically becomes an MCP tool:

# api/mcp/router.py

@router.get("/tools/my_new_tool")
def my_new_tool(param: str):
    """Description of what this tool does (shown to the LLM)."""
    return {"result": "..."}

The function name becomes the tool name, and the docstring becomes the tool description that OpenAI sees. Parameters are extracted from the function signature