MCP Server

🚀 Azure AI Agent Service + Azure AI Search MCP Server

A Model Context Protocol (MCP) server that enables Claude Desktop to search your content via Azure AI services. You can choose to use the Azure AI Agent Service (supporting both document search and web search) or directly integrate it with Azure AI Search.

🚀 Quick Start

This project offers two MCP server implementations to connect Claude Desktop to Azure search capabilities:

1. Azure AI Agent Service Implementation (Recommended) - Utilize the powerful Azure AI Agent Service to provide:

   • Azure AI Search Tool - Search in your indexed documents and obtain enhanced context.
   • Bing Web Content Tool - Fetch real - time information from the Internet.

2. Direct Integration with Azure AI Search - Offers three search modes:

   • Keyword Search - Perform an exact - match search on the content.
   • Vector Similarity Search - Conduct a search based on semantic similarity.
   • Hybrid Search - A comprehensive search combining keywords and context understanding.

✨ Features

• Support for multiple languages and complex queries.
• Real - time web search functionality.
• Scalable toolchain design.
• High - performance document retrieval engine.

🔗 Quick Links

• Azure AI Agent Service Official Documentation
• Azure AI Search Documentation
• Claude Desktop MCP Plugin Development Guide

⚙️ Configuration Requirements

1. Recommended Environment:

   • Windows 10 or higher.
   • Python 3.8+.

2. Required Dependencies:

   • Azure SDK for Python (azure - search - documents, azure - ai - agent).
   • Python network request library (requests).

📦 Installation and Configuration

Using Azure AI Agent Service

1. Create a new .env file and add the following content:

AZURE_AI_AGENT_ENDPOINT=your - endpoint
AZURE_AI_AGENT_KEY=your - key
AZURE_SEARCH_SERVICE_ENDPOINT=your - search - service - endpoint
AZURE_SEARCH_API_KEY=your - search - api - key

2. Install the dependencies:

pip install "mcp[cli]" azure - ai - agent==1.0.0 python - dotenv

3. Start the server:

python azure_ai_agent_service_server.py

Direct Integration with Azure AI Search

1. Create a new .env file and add the following content:

AZURE_SEARCH_SERVICE_ENDPOINT=https://your - service - name.search.windows.net
AZURE_SEARCH_INDEX_NAME=your - index - name
AZURE_SEARCH_API_KEY=your - api - key

2. Install the dependencies:

pip install "mcp[cli]" azure - search - documents==11.5.2 python - dotenv

3. Start the server:

python azure_search_server.py

💻 Usage Examples

- "Search for information about Azure AI."
- "Find the latest developments in large language models."
- "Search for materials related to neural networks using the hybrid mode."

🛠️ Troubleshooting

Server Not Displayed:

1. Check the Claude Desktop logs (located at %APPDATA%\Claude\logs\mcp*.log).
2. Ensure that the file paths and environment variables are configured correctly.
3. Test running the server directly: python azure_ai_agent_service_server.py.

Azure AI Agent Service Issues:

1. Make sure the Azure AI Project is configured correctly.
2. Check if the connection has been created and configured.
3. Verify the status of your Azure account.

⚙️ Customize the Server

• Modify Tool Instructions: Adjust the instructions for each agent to change how it processes queries.
• Add New Tools: Integrate more features using the @mcp.tool() decorator.
• Customize Response Format: Edit how Claude Desktop displays the responses.
• Adjust Web Search Parameters: Optimize the search scope of the Bing search tool.

📄 License

This project is open - sourced under the MIT License.