MCP Servers Collection

A collection of Model Context Protocol (MCP) servers with intelligent clients for various services. This repository demonstrates how to build conversational AI interfaces for different APIs using OpenAI GPT and MCP.

🚀 Featured Servers

📁 Google Drive MCP Server

A comprehensive Google Drive integration with smart discovery, full content access, and intelligent navigation.

Features:

• 🔍 Smart File Discovery - Find files without manual navigation
• 📄 Full Content Access - No limits on document length
• 🧠 Persistent Caching - Files remain accessible for future processing
• 🎯 Intelligent Search - Search across all cached content
• 🗺️ Multi-step Workflows - Handles complex discovery automatically
• 📁 Shared Drive Support - Full Team Drive integration

🍳 Recipe MCP Server

A food and recipe server using the Spoonacular API for discovering and managing recipes.

Features:

• 🔍 Recipe Search - Find recipes by ingredients, cuisine, or dietary restrictions
• 📖 Detailed Instructions - Get complete cooking instructions and ingredients
• 🎲 Random Discovery - Discover new recipes with filtering options
• 💾 Recipe Caching - Save and access previously viewed recipes

📋 Prerequisites

• Python 3.8+
• OpenAI API Key - For GPT integration
• Conda or UV - For package management
• Node.js - For MCP Inspector (testing)

🛠️ Quick Start

1. Clone the Repository

git clone https://github.com/YOUR_ACCOUNT/mcp-servers-collection.git
cd mcp-servers-collection

2. Set Up Environment

# Create conda environment
conda create -n mcp-servers python=3.11
conda activate mcp-servers

# Install base dependencies
pip install mcp openai requests python-dotenv

3. Configure API Keys

Create a .env file in the project root:

OPENAI_API_KEY=your_openai_api_key_here
SPOONACULAR_API_KEY=your_spoonacular_api_key_here

📁 Google Drive Server

Setup

1. Google Cloud Setup:

   • Create a Google Cloud project
   • Enable Google Drive API
   • Create OAuth 2.0 credentials
   • Download credentials.json to project root

2. Install Dependencies:

   pip install google-auth google-auth-oauthlib google-auth-httplib2 google-api-python-client

3. Test the Server:

   cd GDrive
   python gdrive_server.py

4. Run the Client:

   python gdrive_chatbot.py

Example Usage

You: Find the documentation folder in the Team Projects shared drive
🔧 Calling tool: list_shared_drives
🔧 Calling tool: list_files with folder_id from shared drive
🤖 Assistant: Found the documentation folder! Here's what's inside...

You: Search for "budget" in all my cached files
🔧 Calling tool: search_cached_content
🤖 Assistant: Found 3 files containing "budget"...

🍳 Recipe Server

Setup

1. Get Spoonacular API Key:

   • Visit https://spoonacular.com/food-api
   • Sign up for free (150 requests/day)

2. Test the Server:

   cd Recipe
   python recipe_server.py

3. Run the Client:

   python recipe_chatbot.py

Example Usage

You: Find me some vegetarian pasta recipes
🔧 Calling tool: search_recipes
🤖 Assistant: Here are 5 great vegetarian pasta recipes...

You: Get details for recipe ID 12345
🔧 Calling tool: get_recipe_details
🤖 Assistant: Here's the complete recipe with ingredients and instructions...

🧪 Testing with MCP Inspector

Both servers can be tested using the MCP Inspector:

# Test Google Drive Server
npx @modelcontextprotocol/inspector python GDrive/gdrive_server.py

# Test Recipe Server
npx @modelcontextprotocol/inspector python Recipe/recipe_server.py

📚 Documentation

Google Drive Server

• Server Setup Guide
• Client Setup Guide

Recipe Server

• Server Setup Guide
• Client Setup Guide

🏗️ Architecture

MCP Server Pattern

User Query → OpenAI GPT → MCP Client → MCP Server → External API
                ↑                                        ↓
            AI Response ← Tool Results ← Tool Execution ← API Response

Key Components

• MCP Server: Exposes tools and resources via stdio
• MCP Client: Connects to server and integrates with OpenAI
• OpenAI Integration: Provides natural language interface
• Caching Layer: Stores data for future processing

🔧 Available Tools & Resources

Google Drive Server

Tools:

• list_files - List files and folders
• list_shared_drives - List Team Drives
• search_files - Global file search
• get_file_details - File metadata
• read_file - Download and cache content
• create_folder - Create new folders

Resources:

• gdrive://listings - Recent folder explorations
• gdrive://files - Cached file metadata
• gdrive://content/{file_id} - Full file content
• gdrive://search/{query} - Content search results
• gdrive://summary - Cache overview

Prompts:

• explore_shared_drive - Guide shared drive exploration
• find_file_by_name - Help locate specific files
• analyze_folder_structure - Understand organization
• content_discovery - Find content by topic
• setup_workflow - Optimize workflows

Recipe Server

Tools:

• search_recipes - Find recipes by criteria
• get_recipe_details - Complete recipe information
• get_random_recipes - Discover new recipes

Resources:

• recipes://searches - Recent search history
• recipes://saved - Cached recipe details

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Add your MCP server following the established patterns
4. Include comprehensive documentation
5. Submit a pull request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Model Context Protocol - For the MCP framework
• FastMCP - For simplified MCP server development
• OpenAI - For GPT integration
• Spoonacular API - For recipe data
• Google Drive API - For file management

📞 Support

For questions and support:

• Check the documentation in each server's directory
• Review the setup guides for troubleshooting
• Open an issue for bugs or feature requests

---

Happy coding with MCP! 🚀