README

Azure AI Vision Face MCP-Server

Introducing a Face Liveness MCP Server to allow the embedding of proof of presence during Agentic AI workflows.

Preview of Single Tool with Progress

Here's a preview of the MCP server in liveness-server-typescript-tool-with-progress folder, using the “Tool with Progress” feature to consolidate the liveness-check workflow into a single step, removing the need for users to type “Done.” Support for this feature in Claude Desktop is not yet available.

Face Liveness Detection

For more information, visit Face Liveness Detection

Running MCP Server

add dependency

go to liveness-server-typescript folder, run

$ npm install --save-dev typescript @types/node
$ npm install
$ npm run build

then in the build folder, you have index.js

sample for vs code

add .vscode/mcp.json

{
    "servers": {
        "liveness-server": {
            "type": "stdio",
            "command": "node",
            "args": [
                "${workspaceFolder}/liveness-server-typescript/build/stdio.js"
            ],
            "env": {
                "FACEAPI_ENDPOINT": "apiendpoint",
                "FACEAPI_KEY": "apikey",
                "FACEAPI_WEBSITE": "https://liveness-webapp.azurewebsites.net",
                "SESSION_IMAGE_DIR": "${workspaceFolder}/liveness-server-typescript/build/"
            }
        }
    }
}

Sample Claude config should be:

{
  "mcpServers": {
    "liveness-server": {
      "command": "node",
      "args": ["YOUR_PATH/build/stdio.js"],
      "env": {
                "FACEAPI_ENDPOINT": "apiendpoint",
                "FACEAPI_KEY": "apikey",
                "FACEAPI_WEBSITE": "https://yourexample.azurewebsites.net",
                "SESSION_IMAGE_DIR": "D:\somePATH"
        }
    }
  }
}

Localation of session image

If you don't set the variable for SESSION_IMAGE_DIR, images will not be saved.

Liveness with verify

When you have a verify image set, it will switch liveness with verify mode. sample:

"VERIFY_IMAGE_FILE_NAME": "${workspaceFolder}/liveness-server-typescript/build/0b5db043-951c-49d4-9109-e11cb558bb79/sessionImage.jpg",

azure-ai-vision-face-mcp-server FAQ

How do I install the Azure AI Vision Face MCP Server?

Navigate to the 'liveness-server-typescript' folder and run 'npm install' to install dependencies.

What is face liveness detection?

Face liveness detection verifies that a face is real and present, preventing spoofing attacks.

Can this MCP server be used with multiple LLM providers?

Yes, it supports integration with various LLM providers like OpenAI, Claude, and Gemini.

Does the server support progress tracking during liveness checks?

Yes, it uses the 'Tool with Progress' feature to consolidate the workflow into a single step.

Is this MCP server suitable for real-time authentication?

Yes, it is designed for real-time face liveness verification in AI workflows.

Where can I find more information about face liveness detection?

Visit the Microsoft documentation at https://learn.microsoft.com/en-us/azure/ai-services/computer-vision/concept-face-liveness-detection.

What programming language is used for this MCP server?

The server is implemented in TypeScript.

Can this MCP server be integrated into existing AI agent workflows?

Yes, it is designed to embed proof of presence seamlessly into agentic AI workflows.