Google Calendar Assistant

This Python application creates a server that communicates with Google's Gemini AI to interact with your Google Calendar. It uses the Model Context Protocol (MCP) to connect with Google Calendar via SSE (Server-Sent Events).

How It Works

The application follows this flow:

1. Server Initialization:

   • Connects to the MCP server that's already connected to your Google account
   • Retrieves the list of available tools to interact with Google Calendar

2. User Request Processing:

   • User makes a request (e.g., "Tell me what are my meetings tomorrow")
   • The request is sent to Gemini AI along with the available tools list

3. AI Processing:

   • Gemini AI analyzes the request to determine if tools are needed
   • If tools are needed (e.g., for calendar operations), Gemini will use the MCP server to interact with Google Calendar
   • If no tools are needed (e.g., casual conversation), Gemini will respond directly

4. Response Delivery:

   • Gemini sends its analysis/response back to the user
   • The cycle continues with new user requests

Prerequisites

• Python 3.8 or higher
• Gemini API key
• Access to the MCP server connected to your Google Calendar

Installation

1. Clone this repository or download the files
2. Install the required packages:

pip install -r requirements.txt

3. Make sure your .env file contains your Gemini API key and MCP server URL:

GEMINI_API_KEY=your_api_key_here
MCP_SERVER_URL=your_mcp_server_url_here

Usage

Run the application with:

python calendar_assistant.py

In the interactive chat:

• Type your questions or commands about your calendar
• Type refresh to clear the conversation history
• Type quit to exit the application

Example Interactions

Here are some examples of how you can interact with the assistant:

• "Tell me what are my meetings tomorrow"
• "Create a meeting for tomorrow at 5PM with the name 'bank stuff'"
• "What's my schedule for next week?"
• "Cancel my 3PM meeting today"
• "Reschedule my meeting with John to Friday at 10AM"

Successful Tests

The following features have been successfully tested:

• "Show me events for next week/tomorrow/next month"
• "Create an event tomorrow at 3 PM"
• "Delete the event called test tomorrow"

The assistant correctly handles relative dates and local time zones for creating and searching events.

Code Structure

• calendar_assistant.py: Main application file containing the CalendarAssistant class and chat loop
• requirements.txt: List of required Python packages
• .env: Environment variables file for storing the Gemini API key
• logs/: Directory for storing application logs

Technical Details

• MCP (Model Context Protocol): An open standard for connecting AI applications with external tools and data
• Gemini 2.0 Flash: Google's AI model used for processing user requests
• SSE (Server-Sent Events): Protocol used for communication with the MCP server

Troubleshooting

If you encounter issues:

1. Check the logs in the logs/ directory
2. Ensure your Gemini API key is correct
3. Verify that the MCP server URL is accessible

Future Improvements

The following improvements are planned for future versions:

1. Enhanced User Interface:

   • Remove logs from the interface for a cleaner user experience
   • Implement a web interface for easier access

2. New Features:

   • Add voice control for more natural interaction

3. Technical Improvements:

   • Optimize MCP connection management for better stability
   • Improve error handling and user messages
   • Add automated tests to ensure reliability