New Feature: Use Web RTC https://learn.microsoft.com/en-in/azure/ai-services/openai/how-to/realtime-audio-webrtc in Semantic kernel.

[kkdubey]

---

name: Feature request
about: How can i use Web RTC in Semantic kernel.

---

[moonbox3]

Hi @kkdubey, can you give some more context, please? Is there a docs bug? A better explanation needed for realtime audio? Something different? Thank you.

[kkdubey]

@moonbox3 In my application, I'm using Azure CallAutomation service with AI Foundry's real-time model and semantic kernel for multiple agents.

I want to incorporate WebRTC live voice into my service.

[moonbox3]

Hi @kkdubey, we have this ACS call automation demo, but it uses our AzureRealtimeWebsocket instead of AzureRealtimeWebRTC. WebRTC will require extra logic to connect, and for ACS it means one also needs to do conversions because it uses Websocket. We don't have a demo/sample showing this at the moment.

We do have a simpler WebRTC sample here that you may be able to use for inspiration. You can switch out the OpenAIRealtimeWebRTC for AzureRealtimeWebRTC.