AgentQuickstart

Agent Quickstart

For common setup instructions, troubleshooting, and detailed information, see the C# Samples README

This sample demonstrates how to build a real-time voice assistant that connects to an Azure AI Foundry agent. The agent manages model selection, instructions, and tools, enabling sophisticated conversational AI scenarios.

What Makes This Sample Unique

This sample showcases:

• Agent Integration: Routes voice requests to a deployed Azure AI Foundry agent
• Proactive Greeting: Agent initiates the conversation with a welcome message
• Agent Tools: Leverages tools and functions configured in your agent
• Azure Authentication Only: Uses Azure credentials (API key auth not supported for agents)

Prerequisites

• Azure AI Foundry project with a deployed agent
• Azure CLI for authentication
• See C# Samples README for common prerequisites

Quick Start

1. Authenticate with Azure:

   az login

2. Update appsettings.json:

   {
     "VoiceLive": {
       "Endpoint": "https://your-endpoint.services.ai.azure.com/",
       "Voice": "en-US-AvaNeural"
     },
     "Agent": {
       "Id": "asst_your-agent-id",
       "ProjectName": "your-project-name"
     }
   }

3. Run the sample:

   dotnet build
   dotnet run

Command Line Options

# Run with settings from appsettings.json
dotnet run

# Run with command line parameters
dotnet run --agent-id "asst_ABC123" --agent-project-name "my-project"

# Run with custom voice
dotnet run --voice "en-US-JennyNeural" --verbose

Available Options

• --endpoint: Azure VoiceLive endpoint URL
• --agent-id: Azure AI Foundry agent ID (format: asst_xxxxx)
• --agent-project-name: Azure AI Foundry project name
• --voice: Voice for the assistant (default: "en-US-AvaNeural")
• --verbose: Enable detailed logging

See C# Samples README for available voices, troubleshooting, and additional resources.