MCP AI Agent

🚀 MCP Intelligent Agent

This project showcases an intelligent agent capable of performing complex mathematical operations, creating visual presentations, and sending email results via Google services. It demonstrates the ability of multiple MCP (Model Context Protocol) servers to work together to solve composite tasks.

🚀 Quick Start

The main goal of this project is to complete the following task:

Calculate the exponential sum of the first five Fibonacci numbers, write the answer into a slide, download it as a PNG file, and send it via email.

This involves:

1. Mathematical calculations (Fibonacci + Exponentiation)
2. Visual presentation (Google Slides)
3. Communication (Gmail)

✨ Features

System Architecture

The system consists of three main MCP servers:

1. Math Server (mcp_servers/math_server.py)

   • Calculate the Fibonacci sequence
   • Perform exponential operations
   • Handle mathematical calculations

2. Slides Server (mcp_servers/slides_server.py)

   • Create Google Slides presentations
   • Add text to slides
   • Download slides as PNG image files

3. Gmail Server (mcp_servers/gmail_server.py)

   • Send emails with attachments
   • Manage email operations (read, move to spam, mark as read)
   • Handle Gmail API authentication

📦 Installation

Prerequisites

• Python 3.8+
• Linux/Unix environment (tested on Linux 5.15.0 - 136-generic)
• X11 display server (for slide functionality)
• Google account with access to Gmail and Google Slides

Steps

1. Clone the repository.
2. Install dependencies:

pip install -r requirements.txt

Google API Setup

1. Gmail Setup:

   • Create a project in the Google Cloud Console.
   • Enable the Gmail API.
   • Create OAuth 2.0 credentials.
   • Save the credentials as mcp_servers/credentials.json.

2. Environment Variables (optional):

   • GMAIL_CREDS_FILE: Path to the credentials file.
   • GMAIL_TOKEN_FILE: Path to the token file.

💻 Usage Examples

Basic Usage

1. Start the agent:

python agent.py

2. The system will perform the following operations:
   • Calculate the Fibonacci sequence and exponential values.
   • Create a slide containing the results.
   • Save the slide as a PNG file to content/slide.png.
   • Send the PNG file via email.

📚 Documentation

Project Structure

.
├── agent.py              # Main agent implementation file
├── ai.py                 # Gemini interaction
├── prompt.py             # Prompts used in the project
├── mcp_servers/
│   ├── slides_server.py  # Google Slides operations
│   ├── gmail_server.py   # Gmail operations
│   └── math_server.py    # Mathematical calculations
├── content/              # Generated content (slides, images)
└── requirements.txt      # Project dependencies

Troubleshooting

• Display Issues: Ensure that the X11 display server is running.
• Gmail Authorization:
  • You will be prompted for Google account access on the first run.
  • If authorization fails, delete the token.json file and try again.
• File Paths:
  • Images are saved as content/slide.png.
  • Specify file locations using absolute paths.

Example Output

After running the agent, you should see:

1. Mathematical calculation results displayed in the console.
2. A PNG file in the content directory.
3. An email with the PNG attachment received.

Execution Log