Azure AI Vision Face API MCP Server

Introducing a Face Detection and Recognition MCP Server to allow the embedding of face attribute detection and face recognition during Agentic AI workflows.

https://github.com/user-attachments/assets/dac4ef24-2043-47a2-8858-c965970254f9

Face Detection and Recognition API

For more information, visit Face API

Running MCP Server

Installation

1. Using UV

Refer to the Installation Guide for instructions on installing uv.

2. Clone this project

We provide several example images in the example folder. Please clone this project to try the example prompts below.

git clone https://github.com/Azure-Samples/azure-ai-vision-face-api-mcp-server.git

3. Set Up Azure AI Vision Face API

• Go to the Azure Portal and create a new Face resource.
• After deployment, navigate to the resource and copy the Endpoint URL and one of the Key from the "Keys and Endpoint" section.
• You will be prompted to enter the following environment variables the first time you start the MCP server:
  • Azure AI Face API Endpoint: The endpoint URL of your Azure Face API deployment.
  • Azure AI Face API Key: The API key for your Azure Face API resource.

4. (Optional) Deploy GPT-4.1 on Azure OpenAI for Open-Set Attribute Detection

• To enable open-set face attribute detection, you need access to Azure OpenAI.
• In the Azure Portal, create an Azure OpenAI resource and deploy a GPT-4.1 model.
• Once deployed, obtain the Endpoint URL and API Key for your OpenAI resource.
• You will be prompted to enter the following environment variables the first time you start the MCP server:
  • Azure OpenAI Endpoint: The endpoint URL of your Azure OpenAI deployment.
  • Azure OpenAI API Key: The API key for your Azure OpenAI resource.
• For more details about using , see the Azure OpenAI documentation.

5. (Optional) Configure Azure Storage for Image Management

• To enable image management and access, you need an Azure Storage account.
• In the Azure Portal, create an Azure Storage resource and set up a container for your images.
• Once created, obtain your Storage Account Name, Container Name, and a SAS Token for secure access.
• You will be prompted to enter the following environment variables the first time you start the MCP server:
  • AZURE STORAGE ACCOUNT: The name of your Azure Storage account.
  • AZURE STORAGE CONTAINER: The name of your image container.
  • AZURE STORAGE SAS TOKEN: The SAS token for your storage container.
• For more details about using Azure Storage, see the Azure Storage documentation.

6. Interact with our MCP tools using Visual Studio Code GitHub Copilot

• Install GitHub Copilot Visual Studio Code extension.
• Rename the file .vscode/mcp-bp.json to .vscode/mcp.json (or copy .vscode/mcp-bp.json to your existing .vscode/mcp.json) and press Start above the azure_ai_vision_face_api_mcp_server in .vscode/mcp.json.
• Follow this guide to add our MCP tools in the workspace and start a conversation with GitHub Copilot and use MCP tools in agent mode leveraging GPT-4.1.

7. (Optional) Prepare Environment for Local Testing

• Before running pytest or other scripts outside of MCP, copy .env.example to .env

... View full README on GitHub