Edit

Use Azure MCP Server with Azure Speech in Foundry Tools

Azure Model Context Protocol (MCP) Server enables AI assistants like GitHub Copilot, Claude Desktop, and others to interact with Azure Speech in Foundry Tools through natural language commands. This integration allows you to manage text-to-speech and speech-to-text capabilities without writing code or remembering complex CLI syntax.

Azure Speech in Foundry Tools provides speech to text, text to speech, and other capabilities through a Microsoft Foundry resource. You can transcribe speech to text with high accuracy, produce natural-sounding text-to-speech voices, translate spoken audio, and conduct live AI voice conversations.

What is the Azure MCP Server?

The Azure MCP Server enables AI agents and assistants to interact with Azure resources through natural language commands using the Model Context Protocol (MCP). Instead of manually navigating portals or writing scripts, you can describe what you want to accomplish, and the AI assistant uses the Azure MCP Server tools to perform the actions.

For Azure Speech developers, this means you can:

  • Transcribe speech from audio files.
  • Synthesize speech from text input.

Prerequisites

To use the Azure MCP Server with Azure Speech in Foundry, you need:

MCP client requirements

You need an AI assistant or development environment that supports the Model Context Protocol. Choose one:

AI-powered code editors:

Programmatic integration:

For complete setup instructions, see Get started with Azure MCP Server.

Where can you use Azure MCP Server?

The Azure MCP Server works in three primary contexts:

In AI-powered chat and code editors

Use the Azure MCP Server directly within AI assistants and code editors. As you chat about your Azure resources, the AI assistant automatically invokes Azure MCP Server tools to retrieve information, make changes, or answer questions. This is the most common usage pattern.

Get started with:

In programmatic applications

Integrate the Azure MCP Server into your applications using the MCP SDK. Your app acts as an MCP client and invokes Azure MCP Server tools programmatically. This approach is useful for building custom automation, chatbots, or intelligent applications that need Azure integration.

Get started with:

In self-hosted scenarios

Deploy the Azure MCP Server in your own environment for advanced control, security requirements, or custom modifications. You can run it locally, in containers, or integrate it into existing infrastructure. This pattern suits enterprise scenarios requiring air-gapped environments or custom authentication flows.

Learn how to:

Available tools for Azure Speech in Foundry

Azure MCP Server provides the following tools for Azure Speech in Foundry operations:

Tool Description
speech stt recognize Recognize speech from audio files and convert it to text.
speech tts synthesize Generate audio files from text using neural text-to-speech.

For detailed information about each tool, including parameters and examples, see Azure Speech in Foundry tools for Azure MCP Server.

Get started

Ready to use Azure MCP Server with your Azure Speech resources?

  1. Set up your environment: Choose an AI assistant or development tool that supports MCP. For setup and authentication instructions, see the links in the Where can you use Azure MCP Server? section above.

  2. Start exploring: Ask your AI assistant questions about your speech apps or request operations. Try prompts like:

    • "Recognize speech from the audio file 'meeting-recording.mp3'."
    • "Convert 'meeting-recording.wav' audio to text and save it as 'transcript.txt'."
    • "Get transcription results from the audio recording in the project folder."
  3. Learn more: Review the Azure Speech tools reference for all available capabilities and detailed parameter information.

Best practices

  • Use supported audio formats: Ensure you use WAV, MP3, or other supported formats for speech recognition.
  • Specify language for accuracy: Always specify the language parameter to improve recognition accuracy.
  • Choose voice selection wisely: Select appropriate voice options to match the desired tone for synthesized speech.