> ## Documentation Index > Fetch the complete documentation index at: https://hobbyist-e43fa225.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # Quickstart: Voice Agent with Foundry Agent Service (new) > Learn how to create a real-time voice agent with Foundry Agent Service and Voice Live. export const ZonePivot = ({group, options = [], defaultValue, label = "Choose an experience"}) => { const values = options.map(option => option.id); const optionKey = options.map(option => `${option.id}:${option.title}`).join("|"); const [activePivot, setActivePivot] = useState(defaultValue || values[0]); const slugify = value => value.toLowerCase().replace(/[^a-z0-9]+/g, "-").replace(/^-|-$/g, ""); const resolvePivot = () => { if (typeof window === "undefined") return defaultValue || values[0]; const params = new URLSearchParams(window.location.search); const requested = params.get("pivots"); if (requested) { const requestedIds = requested.split(",").map(value => value.trim()).filter(Boolean); const match = requestedIds.find(id => values.includes(id)); if (match) return match; } const hash = window.location.hash.replace(/^#/, ""); if (hash) { const match = options.find(option => option.id === hash || slugify(option.title) === hash); if (match) return match.id; } try { const stored = window.localStorage.getItem(`foundry-zone-pivot:${group}`); if (values.includes(stored)) return stored; } catch { return defaultValue || values[0]; } return defaultValue || values[0]; }; const publishPivotChange = value => { if (typeof window === "undefined") return; window.dispatchEvent(new CustomEvent("foundry-zone-pivot-change", { detail: { group, value } })); }; const syncTableOfContents = () => { if (typeof window === "undefined") return; window.requestAnimationFrame(() => { const toc = document.getElementById("table-of-contents-content"); if (!toc) return; const links = Array.from(toc.querySelectorAll('a[href^="#"]')); for (const link of links) { const item = link.closest("li"); const rawId = link.getAttribute("href")?.slice(1); if (!item || !rawId) continue; let id = rawId; try { id = decodeURIComponent(rawId); } catch {} item.style.display = document.getElementById(id) ? "" : "none"; } }); }; useEffect(() => { const resolvedPivot = resolvePivot(); setActivePivot(resolvedPivot); publishPivotChange(resolvedPivot); window.setTimeout(syncTableOfContents, 0); }, [group, defaultValue, values.join("|"), optionKey]); const selectPivot = value => { setActivePivot(value); if (typeof window !== "undefined") { try { window.localStorage.setItem(`foundry-zone-pivot:${group}`, value); } catch {} const url = new URL(window.location.href); const current = url.searchParams.get("pivots"); const preserved = current ? current.split(",").map(id => id.trim()).filter(id => id && !values.includes(id)) : []; url.searchParams.set("pivots", [...preserved, value].join(",")); window.history.replaceState(null, "", `${url.pathname}${url.search}${url.hash}`); } publishPivotChange(value); window.setTimeout(syncTableOfContents, 0); }; if (options.length < 2) return null; return
{label}
{options.map(option => { const selected = option.id === activePivot; return ; })}
; }; export const ZoneContent = ({group, value, options = [], values = [], defaultValue, children}) => { const optionKey = options.map(option => `${option.id}:${option.title}`).join("|"); const [activePivot, setActivePivot] = useState(defaultValue || values[0]); const slugify = value => value.toLowerCase().replace(/[^a-z0-9]+/g, "-").replace(/^-|-$/g, ""); const resolvePivot = () => { if (typeof window === "undefined") return defaultValue || values[0]; const params = new URLSearchParams(window.location.search); const requested = params.get("pivots"); if (requested) { const requestedIds = requested.split(",").map(value => value.trim()).filter(Boolean); const match = requestedIds.find(id => values.includes(id)); if (match) return match; } const hash = window.location.hash.replace(/^#/, ""); if (hash) { const match = options.find(option => option.id === hash || slugify(option.title) === hash); if (match) return match.id; } try { const stored = window.localStorage.getItem(`foundry-zone-pivot:${group}`); if (values.includes(stored)) return stored; } catch { return defaultValue || values[0]; } return defaultValue || values[0]; }; useEffect(() => { setActivePivot(resolvePivot()); }, [group, defaultValue, values.join("|"), optionKey]); useEffect(() => { const onPivotChange = event => { if (event.detail?.group === group && values.includes(event.detail.value)) { setActivePivot(event.detail.value); } }; window.addEventListener("foundry-zone-pivot-change", onPivotChange); return () => window.removeEventListener("foundry-zone-pivot-change", onPivotChange); }, [group, values.join("|")]); if (activePivot !== value) return null; return <>{children}; }; Learn how to use Voice Live with [Microsoft Foundry Agent Service](https://learn.microsoft.com/azure/ai-foundry/agents/overview) and [Azure Speech in Foundry Tools](https://learn.microsoft.com/azure/ai-services/speech-service/overview) in the Microsoft Foundry portal. You can create and run an application to use Voice Live with agents for real-time voice agents. * Using agents allows leveraging a built-in prompt and configuration managed within the agent itself, rather than specifying instructions in the session code. * Agents encapsulate more complex logic and behaviors, making it easier to manage and update conversational flows without changing the client code. * The agent approach streamlines integration. The agent ID is used to connect and all necessary settings are handled internally, reducing the need for manual configuration in the code. * This separation also supports better maintainability and scalability for scenarios where multiple conversational experiences or business logic variations are needed. To use the Voice Live API without Foundry agents, see the [Voice Live API quickstart](https://learn.microsoft.com/azure/ai-services/speech-service/voice-live-quickstart). To use Voice Live, you don't need to deploy an **audio** model with your Microsoft Foundry resource. Voice Live is fully managed, and the model is automatically deployed for you. For more information about models availability, see the [Voice Live overview documentation](../../../voice-live). ## Prerequisites This document refers to the [Microsoft Foundry (new)](https://learn.microsoft.com/en-us/azure/ai-foundry/what-is-foundry#microsoft-foundry-portals) portal. * An Azure subscription. [Create one for free](https://azure.microsoft.com/pricing/purchase-options/azure-account?cid=msft_learn). * A [Microsoft Foundry resource](../../../../multi-service-resource) created in one of the supported regions. For more information about region availability, see the [Voice Live overview documentation](../../../voice-live). * A Foundry agent created in [Microsoft Foundry](https://ai.azure.com/?cid=learnDocs). For more information about creating an agent, see the [Create an agent quickstart](https://learn.microsoft.com/azure/ai-foundry/agents/quickstart). ## Try out Voice Live in the playground To try out the Voice Live demo, follow these steps: 1. Go to the [Voice Live feature page](https://aka.ms/foundry-voice-live) and select **Open in playground**. 2. Select the agent you created previously to go to the **Agent playground**. 3. Switch the **Voice mode** toggle **On**. Your agent now connects to Voice Live. 4. Expand the right pane, which contains the Voice Live settings. Optionally choose a voice, adjust the VAD settings, set the voice temperature and speed, and change other settings to configure voice behavior. 5. Select **Start session** to start the voice conversation, and select **End** to end the chat session. Learn how to use Voice Live with [Microsoft Foundry Agent Service](https://learn.microsoft.com/azure/ai-foundry/agents/overview) using the VoiceLive SDK for Python. [Reference documentation](https://learn.microsoft.com/python/api/overview/azure/ai-voicelive-readme) | [Package (PyPi)](https://pypi.org/project/azure-ai-voicelive/) | [Additional samples on GitHub](https://aka.ms/voicelive/github-python) You can create and run an application to use Voice Live with agents for real-time voice agents. * Using agents allows leveraging a built-in prompt and configuration managed within the agent itself, rather than specifying instructions in the session code. * Agents encapsulate more complex logic and behaviors, making it easier to manage and update conversational flows without changing the client code. * The agent approach streamlines integration. The agent ID is used to connect and all necessary settings are handled internally, reducing the need for manual configuration in the code. * This separation also supports better maintainability and scalability for scenarios where multiple conversational experiences or business logic variations are needed. To use the Voice Live API without Foundry agents, see the [Voice Live API quickstart](https://learn.microsoft.com/azure/ai-services/speech-service/voice-live-quickstart). To use Voice Live, you don't need to deploy an **audio** model with your Microsoft Foundry resource. Voice Live is fully managed, and the model is automatically deployed for you. For more information about models availability, see the [Voice Live overview documentation](../../../voice-live). Follow the quickstart below or get a fully working web app with browser-based voice UI: ## Prerequisites This document refers to the [Microsoft Foundry (new)](https://learn.microsoft.com/en-us/azure/ai-foundry/what-is-foundry#microsoft-foundry-portals) portal and the latest Foundry Agent Service version. * An Azure subscription. [Create one for free](https://azure.microsoft.com/pricing/purchase-options/azure-account?cid=msft_learn). * Python 3.10 or later version. If you don't have a suitable version of Python installed, you can follow the instructions in the [VS Code Python Tutorial](https://code.visualstudio.com/docs/python/python-tutorial#_install-a-python-interpreter) for the easiest way of installing Python on your operating system. * The required language runtimes, global tools, and Visual Studio Code extensions as described in [Prepare your development environment](https://learn.microsoft.com/en-us/azure/ai-foundry/how-to/develop/install-cli-sdk). * A [Microsoft Foundry resource](../../../../multi-service-resource) created in one of the supported regions. For more information about region availability, see the [Voice Live overview documentation](../../../voice-live). * A model deployed in Microsoft Foundry. If you don't have a model, first complete [Quickstart: Set up Microsoft Foundry resources](https://learn.microsoft.com/en-us/azure/foundry/tutorials/quickstart-create-foundry-resources). - Assign the `Foundry User` role to your user account. You can assign roles in the Azure portal under **Access control (IAM)** > **Add role assignment**. > The Foundry RBAC roles were recently renamed. **Foundry User**, **Foundry Owner**, **Foundry Account Owner**, and **Foundry Project Manager** were previously named Azure AI User, Azure AI Owner, Azure AI Account Owner, and Azure AI Project Manager. You might still see the previous names in some places while the rename rolls out. The role IDs and core permissions are unchanged by the rename. ## Prepare the environment 1. Create a new folder `voice-live-quickstart` and go to the quickstart folder with the following command: ```shell theme={null} mkdir voice-live-quickstart && cd voice-live-quickstart ``` 2. Create a virtual environment. If you already have Python 3.10 or higher installed, you can create a virtual environment using the following commands: # [Windows](#tab/windows) ```bash theme={null} py -3 -m venv .venv .venv\scripts\activate ``` # [Linux](#tab/linux) ```bash theme={null} python3 -m venv .venv source .venv/bin/activate ``` # [macOS](#tab/macos) ```bash theme={null} python3 -m venv .venv source .venv/bin/activate ``` *** Activating the Python environment means that when you run `python` or `pip` from the command line, you then use the Python interpreter contained in the `.venv` folder of your application. You can use the `deactivate` command to exit the python virtual environment, and can later reactivate it when needed. We recommend that you create and activate a new Python environment to use to install the packages you need for this tutorial. Don't install packages into your global Python installation. You should always use a virtual or conda environment when installing Python packages, otherwise you can break your global installation of Python. 1. Create a file named **requirements.txt**. Add the following packages to the file: ```txt theme={null} // Source: requirements.txt (not available) ``` 2. Install the packages: ```bash theme={null} pip install -r requirements.txt ``` ## Retrieve resource information The agent integration requires Entra ID authentication. Key-based authentication isn't supported in Agent mode. Create a new file named `.env` in the folder where you want to run the code. In the `.env` file, add the following environment variables for authentication: ```plaintext theme={null} # Settings for Foundry Agent PROJECT_ENDPOINT= AGENT_NAME="MyVoiceAgent" MODEL_DEPLOYMENT_NAME="gpt-4.1-mini" # Settings for Voice Live AGENT_NAME= # See above AGENT_VERSION= CONVERSATION_ID= PROJECT_NAME= VOICELIVE_ENDPOINT= VOICELIVE_API_VERSION=2026-04-10 ``` Replace the default values with your actual project name, agent name, and endpoint values. | Variable name | Value | | ----------------------------------------- | ---------------------------------------------------------------------------------------------------------------- | | `PROJECT_ENDPOINT` | The Foundry project endpoint copied from the project welcome screen. | | `AGENT_NAME` | The name of the agent to use. | | `AGENT_VERSION` | Optional: The version of the agent to use. | | `CONVERSATION_ID` | Optional: A specific conversation ID to reconnect to. | | `PROJECT_NAME` | The name of your Microsoft Foundry project. Project name is the last element of the project endpoint value. | | `VOICELIVE_ENDPOINT` | This value can be found in the **Keys and Endpoint** section when examining your resource from the Azure portal. | | `FOUNDRY_RESOURCE_OVERRIDE` | Optional: The Foundry resource name hosting the agent project (for example, `my-resource-name`). | | `AGENT_AUTHENTICATION_IDENTITY_CLIENT_ID` | Optional: The managed identity client ID of the Voice Live resource. | Learn more about [keyless authentication](https://learn.microsoft.com/azure/ai-services/authentication) and [setting environment variables](https://learn.microsoft.com/azure/ai-services/cognitive-services-environment-variables). ## Create an agent with Voice Live settings 1. Create a file **create\_agent\_with\_voicelive.py** with the following code: ```python theme={null} // Source: create_agent_v2_with_voicelive.py (not available) ``` 2. Sign in to Azure with the following command: ```shell theme={null} az login ``` 3. Run the Python file. ```shell theme={null} python create_agent_with_voicelive.py ``` ## Talk with a voice agent The sample code in this quickstart uses Microsoft Entra ID for authentication as the current integration only supports this authentication method. The sample connects to Foundry Agent Service by passing `agent_config` in `connect(...)` using these fields: * `agent_name`: The agent name to invoke. * `project_name`: The Foundry project containing the agent. * `agent_version`: Optional pinned version for controlled rollouts. If omitted, the latest version is used. * `conversation_id`: Optional conversation ID to continue prior conversation context. * `foundry_resource_override`: Optional resource name when the agent is hosted on a different Foundry resource. * `authentication_identity_client_id`: Optional managed identity client ID used with cross-resource agent connections. Agent mode in Voice Live doesn't support key-based authentication for agent invocation. Use Microsoft Entra ID (for example, `AzureCliCredential`) for agent access. Voice Live resource configuration might still include API keys for non-agent scenarios. 1. Create the `voice-live-agents-quickstart.py` file with the following code: ```python theme={null} // Source: voice-live-with-agent-v2.py (not available) ``` 2. Sign in to Azure with the following command: ```shell theme={null} az login ``` 3. Run the Python file. ```shell theme={null} python voice-live-agents-quickstart.py ``` 4. You can start speaking with the agent and hear responses. You can interrupt the model by speaking. Enter "Ctrl+C" to quit the conversation. ## Output The output of the script is printed to the console. You see messages indicating the status of the connection, audio stream, and playback. The audio is played back through your speakers or headphones. ```text theme={null} 🎙️ Basic Voice Assistant with Azure VoiceLive SDK ================================================== ============================================================ 🎤 VOICE ASSISTANT READY Start speaking to begin conversation Press Ctrl+C to exit ============================================================ 🎤 Listening... 🤔 Processing... 👤 You said: User Input: Hello. 🎤 Ready for next input... 🤖 Agent responded with audio transcript: Agent Audio Response: Hello! I'm Tobi the agent. How can I assist you today? 🎤 Listening... 🤔 Processing... 👤 You said: User Input: What are the opening hours of the Eiffel Tower? 🎤 Ready for next input... 🤖 Agent responded with audio transcript: Agent Audio Response: The Eiffel Tower's opening hours can vary depending on the season and any special events or maintenance. Generally, the Eiffel Tower is open every day of the year, with the following typical hours: - Mid-June to early September: 9:00 AM to 12:45 AM (last elevator ride up at 12:00 AM) - Rest of the year: 9:30 AM to 11:45 PM (last elevator ride up at 11:00 PM) These times can sometimes change, so it's always best to check the official Eiffel Tower website or contact them directly for the most up-to-date information before your visit. Would you like me to help you find the official website or any other details about visiting the Eiffel Tower? 👋 Voice assistant shut down. Goodbye! ``` The script that you ran creates a log file named `_voicelive.log` in the `logs` folder. ```python theme={null} logging.basicConfig( filename=f'logs/{timestamp}_voicelive.log', filemode="w", format='%(asctime)s:%(name)s:%(levelname)s:%(message)s', level=logging.INFO ) ``` The `voicelive.log` file contains information about the connection to the Voice Live API, including the request and response data. You can view the log file to see the details of the conversation. ```text theme={null} 2026-02-10 18:40:19,183:__main__:INFO:Using Azure token credential 2026-02-10 18:40:19,184:__main__:INFO:Connecting to VoiceLive API with Foundry agent connection MyVoiceAgent for project my-voiceagent-project 2026-02-10 18:40:20,801:azure.identity.aio._internal.decorators:INFO:AzureCliCredential.get_token succeeded 2026-02-10 18:40:21,847:__main__:INFO:AudioProcessor initialized with 24kHz PCM16 mono audio 2026-02-10 18:40:21,847:__main__:INFO:Setting up voice conversation session... 2026-02-10 18:40:21,848:__main__:INFO:Session configuration sent 2026-02-10 18:40:22,174:__main__:INFO:Audio playback system ready 2026-02-10 18:40:22,174:__main__:INFO:Voice assistant ready! Start speaking... 2026-02-10 18:40:22,384:__main__:INFO:Session ready: sess_1m1zrSLJSPjJpzbEOyQpTL 2026-02-10 18:40:22,386:__main__:INFO:Sending proactive greeting request 2026-02-10 18:40:22,419:__main__:INFO:Started audio capture 2026-02-10 18:40:22,722:__main__:INFO:\U0001f916 Assistant response created 2026-02-10 18:40:26,054:__main__:INFO:\U0001f916 Assistant finished speaking 2026-02-10 18:40:26,074:__main__:INFO:\u2705 Response complete 2026-02-10 18:40:32,015:__main__:INFO:User started speaking - stopping playback 2026-02-10 18:40:32,866:__main__:INFO:\U0001f3a4 User stopped speaking 2026-02-10 18:40:32,972:__main__:INFO:\U0001f916 Assistant response created 2026-02-10 18:40:35,750:__main__:INFO:User started speaking - stopping playback 2026-02-10 18:40:35,751:__main__:INFO:\U0001f916 Assistant finished speaking 2026-02-10 18:40:36,171:__main__:INFO:\u2705 Response complete 2026-02-10 18:40:37,117:__main__:INFO:\U0001f3a4 User stopped speaking 2026-02-10 18:40:37,207:__main__:INFO:\U0001f916 Assistant response created 2026-02-10 18:40:41,016:__main__:INFO:\U0001f916 Assistant finished speaking 2026-02-10 18:40:41,023:__main__:INFO:\u2705 Response complete 2026-02-10 18:40:44,818:__main__:INFO:Stopped audio capture 2026-02-10 18:40:44,949:__main__:INFO:Stopped audio playback 2026-02-10 18:40:44,950:__main__:INFO:Audio processor cleaned up ``` Further a session log file is created in the `logs` folder with the name `_conversation.log`. This file contains detailed information about the session, including the request and response data. ```text theme={null} SessionID: sess_1m1zrSLJSPjJpzbEOyQpTL Agent Name: VoiceAgentQuickstartTest Agent Description: Agent ID: None Voice Name: en-US-Ava:DragonHDLatestNeural Voice Type: azure-standard Voice Temperature: 0.8 User Input: Hello. Agent Audio Response: Hello! I'm Tobi the agent. How can I assist you today? User Input: What are the opening hours of the Eiffel Tower? Agent Audio Response: The Eiffel Tower's opening hours can vary depending on the season and any special events or maintenance. Generally, the Eiffel Tower is open every day of the year, with the following typical hours: - Mid-June to early September: 9:00 AM to 12:45 AM (last elevator ride up at 12:00 AM) - Rest of the year: 9:30 AM to 11:45 PM (last elevator ride up at 11:00 PM) These times can sometimes change, so it's always best to check the official Eiffel Tower website or contact them directly for the most up-to-date information before your visit. Would you like me to help you find the official website or any other details about visiting the Eiffel Tower? ``` Here are the key differences between the [technical log](#technical-log) and the [conversation log](#conversation-log): | Aspect | Conversation Log | Technical Log | | ------------------- | --------------------------------- | ------------------------------ | | **Audience** | Business users, content reviewers | Developers, IT operations | | **Content** | What was said in conversations | How the system is working | | **Level** | Application/conversation level | System/infrastructure level | | **Troubleshooting** | "What did the agent say?" | "Why did the connection fail?" | **Example**: If your agent wasn't responding, you'd check: * **voicelive.log** → "WebSocket connection failed" or "Audio stream error" * **conversation.log** → "Did the user actually say anything?" Both logs are complementary - conversation logs for conversation analysis and testing, technical logs for system diagnostics! ### Technical log **Purpose**: Technical debugging and system monitoring **Contents**: * WebSocket connection events * Audio stream status * Error messages and stack traces * System-level events (session.created, response.done, etc.) * Network connectivity issues * Audio processing diagnostics **Format**: Structured logging with timestamps, log levels, and technical details **Use Cases**: * Debugging connection problems * Monitoring system performance * Troubleshooting audio issues * Developer/operations analysis ### Conversation log **Purpose**: Conversation transcript and user experience tracking **Contents**: * Agent and project identification * Session configuration details * **User transcripts**: "Tell me a story", "Stop" * **Agent responses**: Full story text and follow-up responses * Conversation flow and interactions **Format**: Plain text, human-readable conversation format **Use Cases**: * Analyzing conversation quality * Reviewing what was actually said * Understanding user interactions and agent responses * Business/content analysis Learn how to use Voice Live with [Microsoft Foundry Agent Service](https://learn.microsoft.com/azure/ai-foundry/agents/overview) using the VoiceLive SDK for C#. [Reference documentation](https://learn.microsoft.com/dotnet/api/overview/azure/ai.voicelive-readme) | [Package (NuGet)](https://www.nuget.org/packages/Azure.AI.VoiceLive) | [Additional samples on GitHub](https://aka.ms/voicelive/github-csharp) You can create and run an application to use Voice Live with agents for real-time voice agents. * Using agents allows leveraging a built-in prompt and configuration managed within the agent itself, rather than specifying instructions in the session code. * Agents encapsulate more complex logic and behaviors, making it easier to manage and update conversational flows without changing the client code. * The agent approach streamlines integration. The agent ID is used to connect and all necessary settings are handled internally, reducing the need for manual configuration in the code. * This separation also supports better maintainability and scalability for scenarios where multiple conversational experiences or business logic variations are needed. To use the Voice Live API without Foundry agents, see the [Voice Live API quickstart](https://learn.microsoft.com/azure/ai-services/speech-service/voice-live-quickstart). To use Voice Live, you don't need to deploy an **audio** model with your Microsoft Foundry resource. Voice Live is fully managed, and the model is automatically deployed for you. For more information about models availability, see the [Voice Live overview documentation](../../../voice-live). Follow the quickstart below or get a fully working web app with browser-based voice UI: ## Prerequisites This document refers to the [Microsoft Foundry (new)](https://learn.microsoft.com/en-us/azure/ai-foundry/what-is-foundry#microsoft-foundry-portals) portal and the latest Foundry Agent Service version. * An Azure subscription. [Create one for free](https://azure.microsoft.com/pricing/purchase-options/azure-account?cid=msft_learn). * [.NET 8.0 SDK](https://dotnet.microsoft.com/download/dotnet/8.0) or later. * The required language runtimes, global tools, and Visual Studio Code extensions as described in [Prepare your development environment](https://learn.microsoft.com/en-us/azure/ai-foundry/how-to/develop/install-cli-sdk). * A [Microsoft Foundry resource](../../../../multi-service-resource) created in one of the supported regions. For more information about region availability, see the [Voice Live overview documentation](../../../voice-live). * A model deployed in Microsoft Foundry. If you don't have a model, first complete [Quickstart: Set up Microsoft Foundry resources](https://learn.microsoft.com/en-us/azure/foundry/tutorials/quickstart-create-foundry-resources). - Assign the `Foundry User` role to your user account. You can assign roles in the Azure portal under **Access control (IAM)** > **Add role assignment**. > The Foundry RBAC roles were recently renamed. **Foundry User**, **Foundry Owner**, **Foundry Account Owner**, and **Foundry Project Manager** were previously named Azure AI User, Azure AI Owner, Azure AI Account Owner, and Azure AI Project Manager. You might still see the previous names in some places while the rename rolls out. The role IDs and core permissions are unchanged by the rename. ## Prepare the environment 1. Create a new folder `voice-live-quickstart` and go to the quickstart folder with the following command: ```shell theme={null} mkdir voice-live-quickstart && cd voice-live-quickstart ``` 2. Create a **.csproj** file with the following project configuration: ```xml theme={null} // Source: VoiceLiveWithAgent.csproj (not available) ``` 3. Restore NuGet packages: ```shell theme={null} dotnet restore ``` ## Retrieve resource information The agent integration requires Entra ID authentication. Key-based authentication isn't supported in Agent mode. Create a new file named `.env` in the folder where you want to run the code. In the `.env` file, add the following environment variables for authentication: ```plaintext theme={null} # Settings for Foundry Agent PROJECT_ENDPOINT= AGENT_NAME="MyVoiceAgent" MODEL_DEPLOYMENT_NAME="gpt-4.1-mini" # Settings for Voice Live AGENT_NAME= # See above AGENT_VERSION= CONVERSATION_ID= PROJECT_NAME= VOICELIVE_ENDPOINT= VOICELIVE_API_VERSION=2026-04-10 ``` Replace the default values with your actual project name, agent name, and endpoint values. | Variable name | Value | | ----------------------------------------- | ---------------------------------------------------------------------------------------------------------------- | | `PROJECT_ENDPOINT` | The Foundry project endpoint copied from the project welcome screen. | | `AGENT_NAME` | The name of the agent to use. | | `AGENT_VERSION` | Optional: The version of the agent to use. | | `CONVERSATION_ID` | Optional: A specific conversation ID to reconnect to. | | `PROJECT_NAME` | The name of your Microsoft Foundry project. Project name is the last element of the project endpoint value. | | `VOICELIVE_ENDPOINT` | This value can be found in the **Keys and Endpoint** section when examining your resource from the Azure portal. | | `FOUNDRY_RESOURCE_OVERRIDE` | Optional: The Foundry resource name hosting the agent project (for example, `my-resource-name`). | | `AGENT_AUTHENTICATION_IDENTITY_CLIENT_ID` | Optional: The managed identity client ID of the Voice Live resource. | Learn more about [keyless authentication](https://learn.microsoft.com/azure/ai-services/authentication) and [setting environment variables](https://learn.microsoft.com/azure/ai-services/cognitive-services-environment-variables). The C# Foundry Agent SDK (`Azure.AI.Projects`) uses a connection string instead of an endpoint URL. Set `PROJECT_CONNECTION_STRING` to your project connection string (found in the Foundry portal under **Project settings** > **Connected resources**). ## Create an agent with Voice Live settings The agent creation script is a separate utility. Create a temporary console project to run it: 1. Create a separate folder for the agent creation utility: ```shell theme={null} mkdir create-agent && cd create-agent dotnet new console --framework net8.0 ``` 2. Add the required NuGet packages: ```shell theme={null} dotnet add package Azure.AI.Projects dotnet add package Azure.Identity ``` 3. Replace the contents of **Program.cs** with the following code: ```csharp theme={null} // Source: CreateAgentWithVoiceLive.cs (not available) ``` 4. Sign in to Azure with the following command: ```shell theme={null} az login ``` 5. Build and run the agent creation script: ```shell theme={null} dotnet run ``` 6. Return to the quickstart folder: ```shell theme={null} cd .. ``` ## Talk with a voice agent The sample code in this quickstart uses Microsoft Entra ID for authentication as the current integration only supports this authentication method. The sample connects to Foundry Agent Service by passing `AgentSessionConfig` to `StartSessionAsync(SessionTarget.FromAgent(...))` using these properties: * `agentName`: The agent name to invoke. * `projectName`: The Foundry project containing the agent. * `AgentVersion`: Optional pinned version for controlled rollouts. If omitted, the latest version is used. * `ConversationId`: Optional conversation ID to continue prior conversation context. * `FoundryResourceOverride`: Optional resource name when the agent is hosted on a different Foundry resource. * `AuthenticationIdentityClientId`: Optional managed identity client ID used with cross-resource agent connections. Agent mode in Voice Live doesn't support key-based authentication for agent invocation. Use Microsoft Entra ID (for example, `AzureCliCredential`) for agent access. Voice Live resource configuration might still include API keys for non-agent scenarios. 1. Create the **VoiceLiveWithAgentV2.cs** file with the following code: ```csharp theme={null} // Source: VoiceLiveWithAgentV2.cs (not available) ``` 2. Sign in to Azure with the following command: ```shell theme={null} az login ``` 3. Build and run the voice assistant: ```shell theme={null} dotnet run ``` 4. You can start speaking with the agent and hear responses. You can interrupt the model by speaking. Enter "Ctrl+C" to quit the conversation. ## Output The output of the script is printed to the console. You see messages indicating the status of the connection, audio stream, and playback. The audio is played back through your speakers or headphones. ```text theme={null} 🎙️ Basic Foundry Voice Agent with Azure VoiceLive SDK (Agent Mode) ================================================================= ================================================================= 🎤 VOICE ASSISTANT READY Start speaking to begin conversation Press Ctrl+C to exit ================================================================= 🎤 Listening... 🤔 Processing... 👤 You said: Hello. 🎤 Ready for next input... 🤖 Agent responded: Hello! I'm Tobi the agent. How can I assist you today? 🎤 Listening... 🤔 Processing... 👤 You said: What are the opening hours of the Eiffel Tower? 🎤 Ready for next input... 🤖 Agent responded: The Eiffel Tower's opening hours can vary depending on the season and any special events or maintenance. Generally, the Eiffel Tower is open every day of the year, with the following typical hours: - Mid-June to early September: 9:00 AM to 12:45 AM (last elevator ride up at 12:00 AM) - Rest of the year: 9:30 AM to 11:45 PM (last elevator ride up at 11:00 PM) These times can sometimes change, so it's always best to check the official Eiffel Tower website or contact them directly for the most up-to-date information before your visit. Would you like me to help you find the official website or any other details about visiting the Eiffel Tower? 👋 Voice assistant shut down. Goodbye! ``` A conversation log file is created in the `logs` folder with the name `conversation_YYYYMMDD_HHmmss.log`. This file contains session metadata and the conversation transcript, including user inputs and agent responses. ```text theme={null} SessionID: sess_1m1zrSLJSPjJpzbEOyQpTL User Input: Hello. Agent Audio Response: Hello! I'm Tobi the agent. How can I assist you today? User Input: What are the opening hours of the Eiffel Tower? Agent Audio Response: The Eiffel Tower's opening hours can vary depending on the season... ``` Here are the key differences between the [technical log](#technical-log) and the [conversation log](#conversation-log): | Aspect | Conversation Log | Technical Log | | ------------------- | --------------------------------- | ------------------------------ | | **Audience** | Business users, content reviewers | Developers, IT operations | | **Content** | What was said in conversations | How the system is working | | **Level** | Application/conversation level | System/infrastructure level | | **Troubleshooting** | "What did the agent say?" | "Why did the connection fail?" | **Example**: If your agent wasn't responding, you'd check: * **Console log** → "WebSocket connection failed" or "Audio stream error" * **conversation log** → "Did the user actually say anything?" Both logs are complementary - conversation logs for conversation analysis and testing, technical logs for system diagnostics! ### Technical log **Purpose**: Technical debugging and system monitoring **Contents**: * WebSocket connection events * Audio stream status * Error messages and stack traces * System-level events (session.created, response.done, etc.) * Network connectivity issues * Audio processing diagnostics **Format**: Structured logging with timestamps, log levels, and technical details **Use Cases**: * Debugging connection problems * Monitoring system performance * Troubleshooting audio issues * Developer/operations analysis ### Conversation log **Purpose**: Conversation transcript and user experience tracking **Contents**: * Agent and project identification * Session configuration details * **User transcripts**: "Tell me a story", "Stop" * **Agent responses**: Full story text and follow-up responses * Conversation flow and interactions **Format**: Plain text, human-readable conversation format **Use Cases**: * Analyzing conversation quality * Reviewing what was actually said * Understanding user interactions and agent responses * Business/content analysis Learn how to use Voice Live with [Microsoft Foundry Agent Service](https://learn.microsoft.com/azure/ai-foundry/agents/overview) using the VoiceLive SDK for JavaScript. [Reference documentation](https://learn.microsoft.com/javascript/api/overview/azure/ai-voicelive-readme) | [Package (npm)](https://www.npmjs.com/package/@azure/ai-voicelive) | [Additional samples on GitHub](https://aka.ms/voicelive/github-javascript) You can create and run an application to use Voice Live with agents for real-time voice agents. * Using agents allows leveraging a built-in prompt and configuration managed within the agent itself, rather than specifying instructions in the session code. * Agents encapsulate more complex logic and behaviors, making it easier to manage and update conversational flows without changing the client code. * The agent approach streamlines integration. The agent ID is used to connect and all necessary settings are handled internally, reducing the need for manual configuration in the code. * This separation also supports better maintainability and scalability for scenarios where multiple conversational experiences or business logic variations are needed. To use the Voice Live API without Foundry agents, see the [Voice Live API quickstart](https://learn.microsoft.com/azure/ai-services/speech-service/voice-live-quickstart). To use Voice Live, you don't need to deploy an **audio** model with your Microsoft Foundry resource. Voice Live is fully managed, and the model is automatically deployed for you. For more information about models availability, see the [Voice Live overview documentation](../../../voice-live). Follow the quickstart below or get a fully working web app with browser-based voice UI: The JavaScript Voice Live SDK is designed for browser-based applications with built-in WebSocket and Web Audio support. This quickstart uses Node.js with `node-record-lpcm16` and `speaker` for a console experience. ## Prerequisites This document refers to the [Microsoft Foundry (new)](https://learn.microsoft.com/en-us/azure/ai-foundry/what-is-foundry#microsoft-foundry-portals) portal and the latest Foundry Agent Service version. * An Azure subscription. [Create one for free](https://azure.microsoft.com/pricing/purchase-options/azure-account?cid=msft_learn). * [Node.js](https://nodejs.org/) version 18 or later. * [SoX](https://sox.sourceforge.io/) installed on your system (required by `node-record-lpcm16` for microphone capture). * The required language runtimes, global tools, and Visual Studio Code extensions as described in [Prepare your development environment](https://learn.microsoft.com/en-us/azure/ai-foundry/how-to/develop/install-cli-sdk). * A [Microsoft Foundry resource](../../../../multi-service-resource) created in one of the supported regions. For more information about region availability, see the [Voice Live overview documentation](../../../voice-live). * A model deployed in Microsoft Foundry. If you don't have a model, first complete [Quickstart: Set up Microsoft Foundry resources](https://learn.microsoft.com/en-us/azure/foundry/tutorials/quickstart-create-foundry-resources). - Assign the `Foundry User` role to your user account. You can assign roles in the Azure portal under **Access control (IAM)** > **Add role assignment**. > The Foundry RBAC roles were recently renamed. **Foundry User**, **Foundry Owner**, **Foundry Account Owner**, and **Foundry Project Manager** were previously named Azure AI User, Azure AI Owner, Azure AI Account Owner, and Azure AI Project Manager. You might still see the previous names in some places while the rename rolls out. The role IDs and core permissions are unchanged by the rename. ## Prepare the environment 1. Create a new folder `voice-live-quickstart` and go to the quickstart folder with the following command: ```shell theme={null} mkdir voice-live-quickstart && cd voice-live-quickstart ``` 2. Create a **package.json** file with the following content: ```json theme={null} // Source: package.json (not available) ``` 3. Install the dependencies: ```shell theme={null} npm install ``` ## Retrieve resource information The agent integration requires Entra ID authentication. Key-based authentication isn't supported in Agent mode. Create a new file named `.env` in the folder where you want to run the code. In the `.env` file, add the following environment variables for authentication: ```plaintext theme={null} # Settings for Foundry Agent PROJECT_ENDPOINT= AGENT_NAME="MyVoiceAgent" MODEL_DEPLOYMENT_NAME="gpt-4.1-mini" # Settings for Voice Live AGENT_NAME= # See above AGENT_VERSION= CONVERSATION_ID= PROJECT_NAME= VOICELIVE_ENDPOINT= VOICELIVE_API_VERSION=2026-04-10 ``` Replace the default values with your actual project name, agent name, and endpoint values. | Variable name | Value | | ----------------------------------------- | ---------------------------------------------------------------------------------------------------------------- | | `PROJECT_ENDPOINT` | The Foundry project endpoint copied from the project welcome screen. | | `AGENT_NAME` | The name of the agent to use. | | `AGENT_VERSION` | Optional: The version of the agent to use. | | `CONVERSATION_ID` | Optional: A specific conversation ID to reconnect to. | | `PROJECT_NAME` | The name of your Microsoft Foundry project. Project name is the last element of the project endpoint value. | | `VOICELIVE_ENDPOINT` | This value can be found in the **Keys and Endpoint** section when examining your resource from the Azure portal. | | `FOUNDRY_RESOURCE_OVERRIDE` | Optional: The Foundry resource name hosting the agent project (for example, `my-resource-name`). | | `AGENT_AUTHENTICATION_IDENTITY_CLIENT_ID` | Optional: The managed identity client ID of the Voice Live resource. | Learn more about [keyless authentication](https://learn.microsoft.com/azure/ai-services/authentication) and [setting environment variables](https://learn.microsoft.com/azure/ai-services/cognitive-services-environment-variables). ## Create an agent with Voice Live settings 1. Create a file **create-agent-with-voicelive.js** with the following code: ```javascript theme={null} // Source: create-agent-with-voicelive.js (not available) ``` 2. Sign in to Azure with the following command: ```shell theme={null} az login ``` 3. Run the agent creation script: ```shell theme={null} node create-agent-with-voicelive.js ``` ## Talk with a voice agent The sample code in this quickstart uses Microsoft Entra ID for authentication as the current integration only supports this authentication method. The sample connects to Foundry Agent Service by passing an `agent` config object to `client.createSession(...)` using these fields: * `agentName`: The agent name to invoke. * `projectName`: The Foundry project containing the agent. * `agentVersion`: Optional pinned version for controlled rollouts. If omitted, the latest version is used. * `conversationId`: Optional conversation ID to continue prior conversation context. * `foundryResourceOverride`: Optional resource name when the agent is hosted on a different Foundry resource. * `authenticationIdentityClientId`: Optional managed identity client ID used with cross-resource agent connections. Agent mode in Voice Live doesn't support key-based authentication for agent invocation. Use Microsoft Entra ID (for example, `DefaultAzureCredential`) for agent access. Voice Live resource configuration might still include API keys for non-agent scenarios. 1. Create the **voice-live-with-agent.js** file with the following code: ```javascript theme={null} // Source: voice-live-with-agent-v2.js (not available) ``` 2. Sign in to Azure with the following command: ```shell theme={null} az login ``` 3. Run the voice assistant: ```shell theme={null} node voice-live-with-agent.js ``` 4. You can start speaking with the agent and hear responses. You can interrupt the model by speaking. Enter "Ctrl+C" to quit the conversation. ## Output The output of the script is printed to the console. You see messages indicating the status of the connection, audio stream, and playback. The audio is played back through your speakers or headphones. ```text theme={null} 🎙️ Basic Foundry Voice Agent with Azure VoiceLive SDK (Agent Mode) ================================================================= ================================================================= 🎤 VOICE ASSISTANT READY Start speaking to begin conversation Press Ctrl+C to exit ================================================================= 🎤 Listening... 🤔 Processing... 👤 You said: Hello. 🎤 Ready for next input... 🤖 Agent responded with audio transcript: Hello! I'm Tobi the agent. How can I assist you today? 🎤 Listening... 🤔 Processing... 👤 You said: What are the opening hours of the Eiffel Tower? 🎤 Ready for next input... 🤖 Agent responded with audio transcript: The Eiffel Tower's opening hours can vary depending on the season and any special events or maintenance. Generally, the Eiffel Tower is open every day of the year, with the following typical hours: - Mid-June to early September: 9:00 AM to 12:45 AM (last elevator ride up at 12:00 AM) - Rest of the year: 9:30 AM to 11:45 PM (last elevator ride up at 11:00 PM) These times can sometimes change, so it's always best to check the official Eiffel Tower website or contact them directly for the most up-to-date information before your visit. Would you like me to help you find the official website or any other details about visiting the Eiffel Tower? 👋 Voice assistant shut down. Goodbye! ``` A conversation log file is created in the `logs` folder with the name `conversation_YYYYMMDD_HHmmss.log`. This file contains session metadata and the conversation transcript, including user inputs and agent responses. ```text theme={null} SessionID: sess_1m1zrSLJSPjJpzbEOyQpTL Agent Name: VoiceAgentQuickstartTest Agent Description: Agent ID: Voice Name: en-US-Ava:DragonHDLatestNeural Voice Type: azure-standard User Input: Hello. Agent Audio Response: Hello! I'm Tobi the agent. How can I assist you today? User Input: What are the opening hours of the Eiffel Tower? Agent Audio Response: The Eiffel Tower's opening hours can vary depending on the season... ``` Here are the key differences between the [technical log](#technical-log) and the [conversation log](#conversation-log): | Aspect | Conversation Log | Technical Log | | ------------------- | --------------------------------- | ------------------------------ | | **Audience** | Business users, content reviewers | Developers, IT operations | | **Content** | What was said in conversations | How the system is working | | **Level** | Application/conversation level | System/infrastructure level | | **Troubleshooting** | "What did the agent say?" | "Why did the connection fail?" | **Example**: If your agent wasn't responding, you'd check: * **Console log** → "WebSocket connection failed" or "Audio stream error" * **conversation log** → "Did the user actually say anything?" Both logs are complementary - conversation logs for conversation analysis and testing, technical logs for system diagnostics! ### Technical log **Purpose**: Technical debugging and system monitoring **Contents**: * WebSocket connection events * Audio stream status * Error messages and stack traces * System-level events (session.created, response.done, etc.) * Network connectivity issues * Audio processing diagnostics **Format**: Console output with bracketed prefixes (for example, `[session]`, `[audio]`, `[init]`) **Use Cases**: * Debugging connection problems * Monitoring system performance * Troubleshooting audio issues * Developer/operations analysis ### Conversation log **Purpose**: Conversation transcript and user experience tracking **Contents**: * Agent and project identification * Session configuration details * **User transcripts**: "Tell me a story", "Stop" * **Agent responses**: Full story text and follow-up responses * Conversation flow and interactions **Format**: Plain text, human-readable conversation format **Use Cases**: * Analyzing conversation quality * Reviewing what was actually said * Understanding user interactions and agent responses * Business/content analysis Learn how to use Voice Live with [Microsoft Foundry Agent Service](https://learn.microsoft.com/azure/ai-foundry/agents/overview) using the VoiceLive SDK for Java. [Reference documentation](https://learn.microsoft.com/java/api/overview/azure/ai-voicelive-readme) | [Package (Maven)](https://central.sonatype.com/artifact/com.azure/azure-ai-voicelive/overview) | [Additional samples on GitHub](https://aka.ms/voicelive/github-java) You can create and run an application to use Voice Live with agents for real-time voice agents. * Using agents allows leveraging a built-in prompt and configuration managed within the agent itself, rather than specifying instructions in the session code. * Agents encapsulate more complex logic and behaviors, making it easier to manage and update conversational flows without changing the client code. * The agent approach streamlines integration. The agent ID is used to connect and all necessary settings are handled internally, reducing the need for manual configuration in the code. * This separation also supports better maintainability and scalability for scenarios where multiple conversational experiences or business logic variations are needed. To use the Voice Live API without Foundry agents, see the [Voice Live API quickstart](https://learn.microsoft.com/azure/ai-services/speech-service/voice-live-quickstart). To use Voice Live, you don't need to deploy an **audio** model with your Microsoft Foundry resource. Voice Live is fully managed, and the model is automatically deployed for you. For more information about models availability, see the [Voice Live overview documentation](../../../voice-live). Follow the quickstart below or get a fully working web app with browser-based voice UI: ## Prerequisites This document refers to the [Microsoft Foundry (new)](https://learn.microsoft.com/en-us/azure/ai-foundry/what-is-foundry#microsoft-foundry-portals) portal and the latest Foundry Agent Service version. * An Azure subscription. [Create one for free](https://azure.microsoft.com/pricing/purchase-options/azure-account?cid=msft_learn). * [Java Development Kit (JDK)](https://learn.microsoft.com/java/azure/jdk/) version 11 or later. * [Apache Maven](https://maven.apache.org/download.cgi) installed. * The required language runtimes, global tools, and Visual Studio Code extensions as described in [Prepare your development environment](https://learn.microsoft.com/en-us/azure/ai-foundry/how-to/develop/install-cli-sdk). * A [Microsoft Foundry resource](../../../../multi-service-resource) created in one of the supported regions. For more information about region availability, see the [Voice Live overview documentation](../../../voice-live). * A model deployed in Microsoft Foundry. If you don't have a model, first complete [Quickstart: Set up Microsoft Foundry resources](https://learn.microsoft.com/en-us/azure/foundry/tutorials/quickstart-create-foundry-resources). - Assign the `Foundry User` role to your user account. You can assign roles in the Azure portal under **Access control (IAM)** > **Add role assignment**. > The Foundry RBAC roles were recently renamed. **Foundry User**, **Foundry Owner**, **Foundry Account Owner**, and **Foundry Project Manager** were previously named Azure AI User, Azure AI Owner, Azure AI Account Owner, and Azure AI Project Manager. You might still see the previous names in some places while the rename rolls out. The role IDs and core permissions are unchanged by the rename. ## Prepare the environment 1. Create a new folder `voice-live-quickstart` and go to the quickstart folder with the following command: ```shell theme={null} mkdir voice-live-quickstart && cd voice-live-quickstart ``` 2. Create a file named **pom.xml** with the following Maven configuration: ```xml theme={null} // Source: pom-agent.xml (not available) ``` 3. Create the Java source directory structure: # [Windows](#tab/windows) ```shell theme={null} mkdir src\main\java ``` # [Linux](#tab/linux) ```bash theme={null} mkdir -p src/main/java ``` # [macOS](#tab/macos) ```bash theme={null} mkdir -p src/main/java ``` *** 4. Download the Maven dependencies: ```shell theme={null} mvn dependency:resolve ``` ## Retrieve resource information The agent integration requires Entra ID authentication. Key-based authentication isn't supported in Agent mode. Create a new file named `.env` in the folder where you want to run the code. In the `.env` file, add the following environment variables for authentication: ```plaintext theme={null} # Settings for Foundry Agent PROJECT_ENDPOINT= AGENT_NAME="MyVoiceAgent" MODEL_DEPLOYMENT_NAME="gpt-4.1-mini" # Settings for Voice Live AGENT_NAME= # See above AGENT_VERSION= CONVERSATION_ID= PROJECT_NAME= VOICELIVE_ENDPOINT= VOICELIVE_API_VERSION=2026-04-10 ``` Replace the default values with your actual project name, agent name, and endpoint values. | Variable name | Value | | ----------------------------------------- | ---------------------------------------------------------------------------------------------------------------- | | `PROJECT_ENDPOINT` | The Foundry project endpoint copied from the project welcome screen. | | `AGENT_NAME` | The name of the agent to use. | | `AGENT_VERSION` | Optional: The version of the agent to use. | | `CONVERSATION_ID` | Optional: A specific conversation ID to reconnect to. | | `PROJECT_NAME` | The name of your Microsoft Foundry project. Project name is the last element of the project endpoint value. | | `VOICELIVE_ENDPOINT` | This value can be found in the **Keys and Endpoint** section when examining your resource from the Azure portal. | | `FOUNDRY_RESOURCE_OVERRIDE` | Optional: The Foundry resource name hosting the agent project (for example, `my-resource-name`). | | `AGENT_AUTHENTICATION_IDENTITY_CLIENT_ID` | Optional: The managed identity client ID of the Voice Live resource. | Learn more about [keyless authentication](https://learn.microsoft.com/azure/ai-services/authentication) and [setting environment variables](https://learn.microsoft.com/azure/ai-services/cognitive-services-environment-variables). ## Create an agent with Voice Live settings 1. Create a file **src/main/java/CreateAgentWithVoiceLive.java** with the following code: ```java theme={null} // Source: CreateAgentWithVoiceLive.java (not available) ``` 2. Sign in to Azure with the following command: ```shell theme={null} az login ``` 3. Build and run the agent creation script: ```shell theme={null} mvn compile exec:java -Dexec.mainClass="CreateAgentWithVoiceLive" -q ``` ## Talk with a voice agent The sample code in this quickstart uses Microsoft Entra ID for authentication as the current integration only supports this authentication method. The sample connects to Foundry Agent Service by passing `AgentSessionConfig` to `startSession(...)` using these fields: * `agentName`: The agent name to invoke. * `projectName`: The Foundry project containing the agent. * `agentVersion`: Optional pinned version for controlled rollouts. If omitted, the latest version is used. * `conversationId`: Optional conversation ID to continue prior conversation context. * `foundryResourceOverride`: Optional resource name when the agent is hosted on a different Foundry resource. * `authenticationIdentityClientId`: Optional managed identity client ID used with cross-resource agent connections. Agent mode in Voice Live doesn't support key-based authentication for agent invocation. Use Microsoft Entra ID (for example, `AzureCliCredential`) for agent access. Voice Live resource configuration might still include API keys for non-agent scenarios. 1. Create the **src/main/java/VoiceLiveWithAgent.java** file with the following code: ```java theme={null} // Source: VoiceLiveWithAgentV2.java (not available) ``` 2. Sign in to Azure with the following command: ```shell theme={null} az login ``` 3. Build and run the voice assistant: ```shell theme={null} mvn compile exec:java -Dexec.mainClass="VoiceLiveWithAgent" -q ``` 4. You can start speaking with the agent and hear responses. You can interrupt the model by speaking. Enter "Ctrl+C" to quit the conversation. ## Output The output of the script is printed to the console. You see messages indicating the status of the connection, audio stream, and playback. The audio is played back through your speakers or headphones. ```text theme={null} 🎙️ Basic Foundry Voice Agent with Azure VoiceLive SDK (Agent Mode) ================================================================= ============================================================ 🎤 VOICE ASSISTANT READY Start speaking to begin conversation Press Ctrl+C to exit ============================================================ 🎤 Listening... 🤔 Processing... 👤 You said: Hello. 🎤 Ready for next input... 🤖 Agent responded: Hello! I'm Tobi the agent. How can I assist you today? 🎤 Listening... 🤔 Processing... 👤 You said: What are the opening hours of the Eiffel Tower? 🎤 Ready for next input... 🤖 Agent responded: The Eiffel Tower's opening hours can vary depending on the season and any special events or maintenance. Generally, the Eiffel Tower is open every day of the year, with the following typical hours: - Mid-June to early September: 9:00 AM to 12:45 AM (last elevator ride up at 12:00 AM) - Rest of the year: 9:30 AM to 11:45 PM (last elevator ride up at 11:00 PM) These times can sometimes change, so it's always best to check the official Eiffel Tower website or contact them directly for the most up-to-date information before your visit. Would you like me to help you find the official website or any other details about visiting the Eiffel Tower? 👋 Voice assistant shut down. Goodbye! ``` The program uses Java's `java.util.logging` framework for technical logs, which are written to the console (stderr) by default. You can configure a logging properties file to redirect output to a file if needed. ```java theme={null} Logger logger = Logger.getLogger(VoiceLiveWithAgentV2.class.getName()); ``` The console output includes technical information about the connection to the Voice Live API, audio processing, and session events: ```text theme={null} 2026-02-10 18:40:19,183 INFO Using Azure token credential 2026-02-10 18:40:19,184 INFO Connecting to VoiceLive API with agent config... 2026-02-10 18:40:21,847 INFO AudioProcessor initialized with 24kHz PCM16 mono audio 2026-02-10 18:40:21,847 INFO Setting up voice conversation session... 2026-02-10 18:40:21,848 INFO Session configuration sent 2026-02-10 18:40:22,174 INFO Audio playback system ready 2026-02-10 18:40:22,174 INFO Voice assistant ready! Start speaking... 2026-02-10 18:40:22,384 INFO Session ready 2026-02-10 18:40:22,386 INFO Sending proactive greeting request 2026-02-10 18:40:22,419 INFO Started audio capture 2026-02-10 18:40:22,722 INFO 🤖 Assistant response created 2026-02-10 18:40:26,054 INFO 🤖 Assistant finished speaking 2026-02-10 18:40:26,074 INFO ✅ Response complete ``` Further, a conversation log file is created in the `logs` folder with the name `conversation_YYYYMMDD_HHmmss.log`. This file contains session metadata and the conversation transcript, including user inputs and agent responses. ```text theme={null} SessionID: sess_1m1zrSLJSPjJpzbEOyQpTL User Input: Hello. Agent Audio Response: Hello! I'm Tobi the agent. How can I assist you today? User Input: What are the opening hours of the Eiffel Tower? Agent Audio Response: The Eiffel Tower's opening hours can vary depending on the season... ``` Here are the key differences between the [technical log](#technical-log) and the [conversation log](#conversation-log): | Aspect | Conversation Log | Technical Log | | ------------------- | --------------------------------- | ------------------------------ | | **Audience** | Business users, content reviewers | Developers, IT operations | | **Content** | What was said in conversations | How the system is working | | **Level** | Application/conversation level | System/infrastructure level | | **Troubleshooting** | "What did the agent say?" | "Why did the connection fail?" | **Example**: If your agent wasn't responding, you'd check: * **Console log** → "WebSocket connection failed" or "Audio stream error" * **conversation log** → "Did the user actually say anything?" Both logs are complementary - conversation logs for conversation analysis and testing, technical logs for system diagnostics! ### Technical log **Purpose**: Technical debugging and system monitoring **Contents**: * WebSocket connection events * Audio stream status * Error messages and stack traces * System-level events (session.created, response.done, etc.) * Network connectivity issues * Audio processing diagnostics **Format**: Structured logging with timestamps, log levels, and technical details **Use Cases**: * Debugging connection problems * Monitoring system performance * Troubleshooting audio issues * Developer/operations analysis ### Conversation log **Purpose**: Conversation transcript and user experience tracking **Contents**: * Agent and project identification * Session configuration details * **User transcripts**: "Tell me a story", "Stop" * **Agent responses**: Full story text and follow-up responses * Conversation flow and interactions **Format**: Plain text, human-readable conversation format **Use Cases**: * Analyzing conversation quality * Reviewing what was actually said * Understanding user interactions and agent responses * Business/content analysis ## Related content * Learn more about [How to build voice agents](/tools-and-knowledge/how-to-voice-agent-integration) * Learn more about [How to use the Voice Live API](./voice-live-how-to) * See the [Voice Live API reference](./voice-live-api-reference-2026-04-10)