This repository contains sample code and resources for working with Azure AI Speech Service Voice Live.
Voice Live enables real-time voice interactions using Azure AI Speech Service. These samples demonstrate how to integrate Voice Live into your applications for various scenarios including conversational AI, voice assistants, and interactive voice experiences.
- Real-time voice interaction samples
- Integration examples with Azure AI Speech Service
- Best practices for Voice Live implementation
- Sample code for common use cases
- Azure subscription - Create one for free
- AI Foundry resource
- Basic knowledge of your preferred programming language
-
Clone this repository:
git clone https://github.com/azure-ai-foundry/voicelive-samples.git cd voicelive-samples -
Follow the instructions in individual sample directories for specific setup requirements.
- Choose your language: Select from C#, Python, or other available languages
- Follow language-specific setup: Each language folder has detailed setup instructions
- Configure credentials: Set up your Azure resources and authentication
- Run a sample: Start with a quickstart sample to see Voice Live in action
For detailed quickstart guides, see the README in your chosen language folder.
This repository contains samples in multiple programming languages. Choose your preferred language to get started:
Complete C# samples demonstrating:
- Agent Quickstart: Connect to Azure AI Foundry agents with proactive greetings
- Model Quickstart: Direct VoiceLive model integration
- Bring-Your-Own-Model (BYOM) Quickstart: Use your own models hosted in Foundry with proactive greetings
- Customer Service Bot: Advanced function calling for customer service scenarios and proactive greetings
- Built with .NET 9.0 and self-contained code
Python samples showcasing:
- Agent Quickstart: Azure AI Foundry agent integration with proactive greetings
- Model Quickstart: Direct model access with flexible authentication
- Bring-Your-Own-Model (BYOM) Quickstart: Use your own models hosted in Foundry with proactive greetings
- Function Calling: Advanced tool integration with custom functions and proactive greetings
- RAG-enabled Voice Assistant: Full-stack voice assistant with Azure AI Search integration and
azddeployment - Built with Python 3.8+ and async/await patterns
JavaScript/TypeScript samples showcasing:
- Basic Web Voice Assistant: Browser-based voice assistant with real-time streaming and barge-in support
- Voice Live Avatar: Avatar-enabled voice conversations with Docker deployment
- Voice Live Car Demo: Voice-Enabled Car Assistant powered by multiple architectures
- Voice Live Interpreter: Real-time speech translation, speech in and speech out
- Voice Live Trader: Real-time trading assistant for stock fund crypto FX trading app
- Built with TypeScript and Web Audio API
Java samples showcasing:
- Model Quickstart: Direct model access with flexible authentication
- Built with Java 11+ and Maven
Each language folder contains detailed setup instructions, configuration examples, and troubleshooting guides specific to that language and platform.
We welcome contributions! Please see our Contributing Guidelines for details.
Please note that this project follows the Microsoft Open Source Code of Conduct.
This project is licensed under the MIT License - see the LICENSE file for details.
Copyright (c) Microsoft Corporation. All rights reserved.
This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft trademarks or logos is subject to and must follow Microsoft's Trademark & Brand Guidelines. Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. Any use of third-party trademarks or logos are subject to those third-party's policies.