WebRTC Live

In the ongoing debate around WebRTC vs. MoQ, the most practical answer may be, “why not both?” The IETF is now examining whether QUIC offers real-time media applications any advantage over WebRTC, with WebTransport and MoQ positioned as possible foundations for the next generation of real-time protocols.

In this episode of WebRTC Live, we welcome back Lorenzo Miniero, Chairman at Meetecho and author of the Janus WebRTC Server. Lorenzo has been experimenting with MoQ and QUIC for several years, looking at MoQ from the perspective of a WebRTC developer. He’ll talk about building the imquic library and adding QUIC-based protocols to Janus, including RTP over QUIC (RoQ) and MoQ. We’ll discuss the problems he ran into integrating these protocols with an existing WebRTC server, the gaps between WebRTC and QUIC-based frameworks, and the right use cases for WebRTC, MoQ, or both.

The episode includes our monthly WebRTC Industry Update with Tsahi Levent-Levi. This month’s topic: Deepgram Acquires Daily: Was It a Bet on Pipecat?

Voice AI applications need a way to store conversation data that supports observability, auditing, and long-term retention. The audio recording or a text transcript alone leaves out critical context. Additional details matter too, such as who was in the conversation, what devices were used, what was said, and how decisions were made.

vCons act as the “memory” of a conversation. They are structured, portable, machine-readable objects that capture conversation data across voice, video, messaging, and email, regardless of platform. This matters especially for Voice AI in regulated industries, like healthtech or fintech, where each conversation needs to be fully auditable.

Our guest is Jeff Pulver, VoIP industry pioneer and founder of the vCon Foundation. We cover the need for vCons, the current status of the IETF draft, where adoption stands, use cases and examples, how AI is changing internet communications, and more. 

The episode include our monthly WebRTC Industry Update with Tsahi Levent-Levi. This month's topic: Chrome's Neural Echo Cancellation Leaves Edge, Electron, and Native Apps Behind.

Human speech is often unstructured and messy. To keep up with real conversation, Voice AI bots must parse partial sentences, filler words, and interruptions, all without the visual cues a human listener relies on. On top of that linguistic challenge, real-world latency budgets and network conditions test every Voice AI pipeline, and users don’t forgive one that stutters.

In this episode of WebRTC Live, host Arin Sime talks with Dr. Varun Singh, Chief Product Technology Officer at Daily.co / Pipecat.ai, about the engineering decisions that separate voice agents that succeed from ones that get abandoned. Topics include where Voice AI is working today and where it isn’t, from well-defined tasks like collections to the open-ended conversations everyone is still chasing, and how Pipecat approaches the problem: composing voice agents from STT, LLM, TTS, and everything in between, fitting in speech-to-speech models, and getting from prototype to production. Also discussed: why WebRTC remains the right transport for real-world networks, the production realities of observability, cost, and failure modes, how to weigh build-vs-buy, and where voice-native UIs are headed.

The episode includes our monthly WebRTC Industry Update with Tsahi Levent-Levi. This month’s topic: OpenAI's GPT-Live Voice AI Architecture and the WARP Protocol.

How can AI scale clinical roleplay training past what live sessions with human trainers can deliver? CETA Global, a nonprofit that trains frontline mental health providers worldwide, built an AI “flight simulator.” On it, psychologists practice difficult client sessions with an AI patient shown as a lip-synced streaming avatar, get live coaching, and replay feedback afterward. Six AI agents work together to play the patient, coach the trainee, and evaluate the session against evidence-based criteria in real time.

Host Arin Sime talks with Andrés Rincon, the WebRTC.ventures lead developer on EBT-Sim. The team took it from concept to working platform in eight weeks for a Google.org Accelerator: Generative AI demo.

They cover how the team made concurrent agents feel like one conversation: a WebSocket-based real-time layer, Gemini agents orchestrated through Google ADK, low-latency speech-to-text and text-to-speech, and  streaming avatars. Andrés explains why the team chose push-to-talk over automatic turn detection and how they built adaptive resistance into the AI client.

The episode includes our monthly WebRTC Industry Update with Tsahi Levent-Levi. This month’s topic: What's the difference between STUN, TURN & ICE servers?.

“Real-time voice AI only works when infrastructure makes latency feel invisible. For OpenAI, that meant changing the shape of our WebRTC deployment without changing what clients expect from WebRTC itself.”

Yi Zhang (LinkedIn), co-author of How OpenAI delivers low-latency voice AI at scale, explains how his team cut pauses, interruptions, and delayed barge-in for ChatGPT voice users and voice agents. He covers running WebRTC inside Kubernetes, separating stateless packet forwarding from stateful session ownership, using ICE ufrag routing to send the first packet to the right server, and what this architecture means for developers building on the OpenAI Realtime API. Yi previously worked at Meta and Microsoft and holds a PhD in Engineering from the University of Wisconsin–Madison.

Also in this episode: our monthly WebRTC Industry Analysis with Tsahi Levent-Levi on where AI tooling fits in WebRTC development.

WhatsApp has over two billion active users, but for developers building real-time communications applications, it’s long been a walled garden. That’s changing. WhatsApp Business Calling is opening new doors for SIP-based integrations, and the technical landscape around it is evolving fast.

In this episode, Dan Jenkins joins us to break down what combining WhatsApp Business Calling with WebRTC and SIP actually looks like in practice: the use cases, the architecture decisions, and the technical challenges that come with it. Dan Jenkins is founder and CEO of NimbleApe, Everycast Labs, and the organizer of the CommCon conference series. As always, Dan will bring his expertise in low latency media, video streaming and broadcast technologies to our conversation.

The episode include our monthly WebRTC Industry Update with Tsahi Levent-Levi. This month's topic: OpenAI’s Secret WebRTC Architecture: No SFU? No Problem!

Half of the call-quality tickets Balázs Kreith investigates trace back to connectivity problems before media quality ever comes into play. A one-hour meeting with five bad minutes at the start can leave users satisfied, while an aggregated quality-of-service score labels the same call as bad.

Balázs explains how QoS and QoE diverge, why a single score misrepresents what users perceive over the course of a call, and why observability needs to be built in from the start so teams can trace problems after deployment. He is a Senior Software Engineer at Riverside.fm, lead developer of the open source ObserveRTC project, and previously worked on WebRTC teams at Whereby and callstats.io.

Also in this episode: Arin Sime and Tsahi Levent-Levi’s monthly WebRTC Industry Chat on the rise of the WebRTC micro app.