Azure Content Understanding MCP Server

This project provides a Model Context Protocol (MCP) server for integrating Azure Content Understanding capabilities with AI systems. The server allows content analysis tools to be exposed through a standard MCP interface.

Project Structure

• ContentUnderstanding.MCP.Server.Stdio: Core MCP server implementation using standard I/O for communication
• ContentUnderstanding.MCP.Tools: Library containing tools for content analysis and Azure - Content Understanding API integration

Features

• Document Analysis: Analyze various file formats using Azure Content Understanding service
• Multiple Analyzer Support: Retrieve and use different analyzers based on content type or your data requirements.

Getting Started

Prerequisites

• .NET 9.0 SDK
• Azure subscription with Content Understanding service
• Azure Blob Storage account

Tools and Components

Content Understanding Client

The ContentUnderstandingClient provides methods for interacting with Azure Content Understanding:

• Create and update analyzers
• List available analyzers
• Submit documents for analysis
• Retrieve analysis results

Document Analysis

The AnalyzeDocument class handles the end-to-end process:

1. Upload the document to Azure Blob Storage
2. Submit for analysis via Content Understanding API
3. Poll for analysis completion
4. Return structured results
5. Clean up temporary blob storage

Sample MCP Configuration

"contentUnderstanding": {
    "command": "dotnet",
    "args": [
        "run",
        "--project",
        "FULL_PATH_TO_PROJECT_FOLDER",
        "--no-build"
    ],
    "env": {
        "ENDPOINT": "CONTENT_UMDERSTANDING_ENDPOINT",
        "API_KEY": "CONTENT_UMDERSTANDING_API_KEY",
        "API_VERSION": "CONTENT_UMDERSTANDING_VERSION",
        "STORAGE_CONTAINER_URL": "STORAGE_CONTAINER_URL",
        "ALLOWED_DIRECTORIES": "SEMICOLON_SEPERATED_LIST_OF_FOLDERS"
    }
}