Azure OpenAI’s Realtime API Now Supports WebRTC: Low-Latency Voice Agents Just Got a Lot Easier

TL;DR: The Realtime API (preview) now supports WebRTC, enabling real-time audio streaming and low-latency interactions, ideal for applications requiring immediate feedback, such as live customer support or interactive voice assistants. This is a game-changer for .NET and Azure developers building voice-first AI agents without custom infrastructure.

Why This Matters Right Now

If you’ve been eyeing voice-based AI agents but got spooked by the complexity of managing WebRTC connections, SIP trunks, and low-latency streaming—breathe. The Realtime API (preview) now supports WebRTC, enabling real-time audio streaming and low-latency interactions. This feature is ideal for applications requiring immediate feedback, such as live customer support or interactive voice assistants. For the past few months, Azure developers had to choose between building voice agents with raw socket management or accepting higher latency. Now, the Realtime API now supports SIP, enabling telephony connections to realtimeapi. Combined with WebRTC support, you’ve got both the modern web path and the legacy telephony path baked in.

What This Unlocks for Your .NET Stack

Real-Time Agent Orchestration

You can now build and deploy AI agents end-to-end in Visual Studio Code with help from GitHub Copilot. AI Toolkit for VS Code lets developers explore models and build agents where they code—with evaluation and tracing in one place. WebRTC support means your voice agents no longer need a separate media gateway.

Cost & Latency Wins

These models are lightweight and highly optimized, delivering real-time voice interaction and audio generation with minimal resource requirements. Their streamlined architecture enables rapid inference and low latency, making them ideal for scenarios where speed and responsiveness are critical—such as voice-based chatbots, real-time translation, and dynamic audio content creation.

Integration Path: Getting Started

If you’re on .NET, here’s the practical next step:

using Azure.AI.OpenAI;

var client = new AzureOpenAIClient(
    new Uri("https://<your-resource>.openai.azure.com/"),
    new DefaultAzureCredential()
);

// The Realtime API client now handles WebRTC negotiation internally
var realtimeClient = client.GetRealtimeClient("gpt-4o-realtime-preview");

// Your voice stream connects via WebRTC—no custom SDP parsing required
await realtimeClient.StartAudioStreamAsync(audioInput);

Starting in August 2025, you can now opt in to the next generation v1 Azure OpenAI APIs which add support for ongoing access to the latest features with no need to specify new api-versions each month. The v1 API removes this dependency, by adding automatic token refresh support to the OpenAI() client.

The Broader Context: Agentic AI on Azure

This WebRTC support doesn’t exist in a vacuum. Azure MCP Server is now generally available, giving your agents the power of cloud and redefining how developers interact with Azure. Built on Model Context Protocol (MCP), it can create a secure, standards-based bridge between Azure services—like AKS, ACA, App Service, Cosmos DB, SQL, AI Foundry, and Fabric—and AI-powered tools such as GitHub Copilot. Imagine managing cloud resources, generating infrastructure-as-code, and troubleshooting deployments—all through natural language, right from your favorite IDE or MCP-compatible client, and all aligned with Azure best practices. For voice agents specifically, Foundry Observability is now in preview, giving developers end-to-end monitoring, built-in metrics, and detailed trace logs of the reasoning steps and tool calls made by agents. This means you can debug a stuck voice interaction without guessing where the latency spike came from.

The Catch: Still Preview, But Production-Ready

WebRTC support is in preview, so don’t ship it to your most critical call center on day one. But according to McKinsey’s 2025 Global AI Trust Survey, the number one barrier to AI adoption is lack of governance and risk-management tools. Task adherence, prompt shields with spotlighting, and PII detection are being put in public preview. These capabilities are built into Azure AI Foundry, helping organizations build with confidence and comply with internal and external standards.

Next Steps

Upgrade your Azure OpenAI SDK to the latest version supporting v1 APIs.
Test WebRTC negotiation in a dev environment using the Realtime API preview.
Integrate with Azure AI Foundry for observability and multi-agent orchestration if you’re building beyond single-agent voice.
Plan your SIP fallback for legacy phone systems—it’s there if you need it.

Azure OpenAI's Realtime API Now Supports WebRTC: Low-Latency Voice Agents Just Got a Lot Easier