ofp-ai-bootcamp2025-trtran by Azure - MCP Server

Microsoft WhoIs MCP Server

This project implements a Model Context Protocol (MCP) Server in Python using the FastMCP library. The server exposes tools for searching person information using the Microsoft Graph API with Azure CLI authentication.

Features

🔍 Person Search: Search for people by name using Microsoft Graph API
📧 Contact Information: Retrieve email addresses, phone numbers, job titles, and aliases
🏢 Organization Details: Get company, department, and office location information
🔐 Azure CLI Authentication: Seamless authentication using Azure CLI credentials
🛡️ Error Handling: Comprehensive error handling and validation
🚀 FastMCP Integration: Built with the latest FastMCP library for optimal performance

API Endpoint

The server queries the Microsoft Graph API endpoint:

GET https://graph.microsoft.com/v1.0/users

Installation

Prerequisites

Python 3.12
Azure CLI installed and authenticated (az login)
Conda (recommended) or pip

Using Conda (Recommended)

Clone the repository:

git clone <repository-url>
cd ofp-ai-bootcamp2025-trtran

Create and activate the conda environment:

conda env create -f environment.yml
conda activate mcp_server_env

Using pip

Install dependencies:
```
pip install -r requirements.txt
```

Configuration

Authentication Setup

The server uses Azure CLI authentication. No manual token configuration is required.

Install Azure CLI if not already installed:
- Windows: Download from Azure CLI releases
- macOS: brew install azure-cli
- Linux: Follow Azure CLI installation guide
Authenticate with Azure:
```
az login
```
Verify authentication:
```
az account show
```

The server will automatically obtain and cache access tokens using your Azure CLI credentials.

Usage

Running the MCP Server

Method 1: Direct Python execution

python mcp_server.py

Method 2: Using the provided batch script (Windows)

start_server.bat

Testing the Server

Run the test suite to verify everything is working:

python test_server.py

Available Tools

1. `search_person`

Search for people by name using Microsoft Graph API.

Parameters:

name (string): The name of the person to search for

Example:

{
  "name": "Tri Tran"
}

Response:

{
  "results": [
    {
      "display_name": "Tri Tran",
      "email": "trtran@microsoft.com",
      "alias": "trtran",
      "job_title": "PRINCIPAL SOFTWARE ENGINEER",
      "office_location": "STUDIO X/1310",
      "phone": "+1 (425) 5382112"
    }
  ],
  "count": 1,
  "total_count": 1,
  "error": null
}

2. Person Information Fields

Each person result includes:

display_name: Full name as it appears in Microsoft's directory
email: Primary email address
alias: Microsoft alias (username without @microsoft.com)
job_title: Current job title
office_location: Office location/building information
phone: Business phone number (if available)

Using with MCP Clients

This server is compatible with any MCP client. Here are some examples:

Claude Desktop (Anthropic)

Add to your Claude Desktop configuration:

{
  "mcpServers": {
    "microsoft-whois": {
      "command": "python",
      "args": ["path/to/mcp_server.py"]
    }
  }
}

Custom Client

import asyncio
from fastmcp import Client

async def search_example():
    client = Client("mcp_server.py")
    
    async with client:
        result = await client.call_tool("search_person", {"name": "Tri Tran"})
        print(result)

asyncio.run(search_example())

Architecture

┌─────────────────┐    ┌─────────────────┐    ┌──────────────────┐
│   MCP Client    │◄──►│  FastMCP Server │◄──►│ Microsoft Graph  │
│  (Claude, etc.) │    │   (mcp_server)  │    │    API + Azure   │
│                 │    │                 │    │  CLI Credentials │
└─────────────────┘    └─────────────────┘    └──────────────────┘

Project Structure

ofp-ai-bootcamp2025-trtran/
├── mcp_server.py          # Main MCP server implementation
├── test_server.py         # Test suite for the server
├── start_server.bat       # Windows batch script to start server
├── setup.bat              # Windows batch script for setup
├── requirements.txt       # Python dependencies
├── environment.yml        # Conda environment configuration
├── server_config.json     # Server configuration
├── README.md             # This file
└── LICENSE               # License file

Dependencies

fastmcp: Latest MCP server framework
httpx: Modern HTTP client for API requests
azure-identity: Azure authentication library for CLI credentials

Error Handling

The server includes comprehensive error handling for:

Missing or invalid Azure CLI authentication
Network timeouts and connectivity issues
API rate limiting and quota errors
Invalid input parameters
Unexpected API responses

Security Considerations

� Uses Azure CLI credentials (no hardcoded tokens)
🛡️ Input validation prevents injection attacks
⏰ Request timeouts prevent hanging connections
📝 Detailed error logging without exposing sensitive information
🔒 Token caching for efficient authentication

Development

Running Tests

python test_server.py

Code Structure

GraphAPIClient: Handles Microsoft Graph API interactions with Azure CLI authentication
search_person: Main tool for person search functionality
Token caching mechanism for improved performance

Troubleshooting

Common Issues

"No access token available. Run 'az login' to authenticate"
- Run az login to authenticate with Azure CLI
- Verify authentication with az account show
"Import 'fastmcp' could not be resolved"
- Install dependencies: pip install -r requirements.txt
"Unauthorized" (401 error)
- Re-authenticate with Azure CLI: az login
- Ensure you have access to Microsoft Graph API
"Request timeout"
- Check your internet connection
- Verify Microsoft Graph API endpoint is accessible

Contributing

Fork the repository
Create a feature branch
Make your changes
Add tests for new functionality
Submit a pull request

License

See the file for details.

Support

For issues and questions:

Check the troubleshooting section above
Review Microsoft Graph API documentation
Check FastMCP documentation at gofastmcp.com
Open an issue in this repository