🧠 MCP Server Project - Multi-Agent AI Tools

This project demonstrates how to build and serve multiple AI tools using the FastMCP framework. It includes a combination of COM-based email automation, external API integration (OpenWeather), local file-based notes, and utility functions—all accessible via an MCP-compatible agent interface.

🔧 Requirements

• Python 3.10+

• Windows (required for Outlook COM)

• .env file with:

  OPENWEATHER_API_KEY=your_openweather_key_here

📁 Directory Structure

. ├── outlook_assistant.py # Outlook email read/reply/compose/search tools ├── weather_mcp.py # Current, forecast, historical, and air quality weather tools ├── sticky_notes.py # Notes tools backed by a text file ├── addition_server.py # Simple addition and greeting tool ├── notes.txt # File to store sticky notes └── README.md # This file

📬 Outlook Email Assistant

A COM-based email assistant that integrates with Microsoft Outlook to:

• List all folders in Outlook

• Fetch and display recent or matching emails

• Read full email contents

• Send replies or compose new emails

🔧 Tools Exposed:

• list_folders()

• list_recent_emails(days, folder_name)

• search_emails(search_term, days, folder_name)

• get_email_by_number(email_number)

• reply_to_email_by_number(email_number, reply_text)

• compose_email(recipient_email, subject, body, cc_email)

⚠️ Must be run on Windows with Microsoft Outlook installed.

🌍 Weather Tools (OpenWeather API)

Provides weather details using OpenWeatherMap API including:

• Current weather

• 5-day forecast

• Air quality

• Weather alerts

• Historical weather (past 5 days)

🌦 Tools Exposed:

• get_current_weather(location)

• get_weather_forecast(location, days)

• get_air_quality(location)

• get_historical_weather(location, date)

• get_weather_alerts(location)

• search_location(query)

🌐 Requires setting OPENWEATHER_API_KEY in .env

📝 AI Sticky Notes

Simple note-taking assistant that stores notes locally in a notes.txt file.

🛠 Tools and Resources:

• add_note(message)

• read_notes()

• get_latest_note() – as resource notes://latest

• note_summary_prompt() – generates a summarization prompt from all notes

➕ Addition Utility

Example of a simple arithmetic and dynamic greeting tool using MCP:

• add(a, b)

• get_greeting(name) → resource greeting://{name}

🏁 Running Any MCP Server

Each file defines its own FastMCP server. Run them individually:

python outlook_assistant.py python weather_mcp.py python sticky_notes.py python addition_server.py

Each server will start on its own FastMCP interface for integration with agents.

🧪 Example Agents

You can now connect these MCP tools with agents (CrewAI, LangChain, etc.) to enable conversational workflows like:

• "What’s the weather in London today?"

• "Summarize my latest sticky notes"

• "Reply to the last unread email from Ajay"

• "What is 42 + 11?"