Private Workshop
WebRTC for Voice AI, live for your team
Voice AI runs on WebRTC, and most teams run it on default, untuned WebRTC built for human calls. This is a 2-hour technical deep dive into what is actually happening on the wire in 2026: what the leading players do, and the specific tuning levers that separate finely optimized services from naïve implementations. Delivered live, for your team, by the two people who track this layer for a living.
Who it is for
Engineering teams adding Voice AI to an existing WebRTC product, now being asked to make it sound right in production.
Voice AI teams running on LiveKit Agents, Pipecat, or a direct stack, who need the WebRTC layer underneath to hold up.
The lead, engineering manager, or architect who owns the real-time stack and the roadmap calls that ride on it.
If you own this stack, this is your opportunity to pressure-test your architecture decision while it is still cheap to change.
It assumes WebRTC familiarity and goes deep fast.
What we cover
The map
What is actually happening on the wire
- The transport choice. WebRTC vs WebSocket vs WebTransport for voice: bitrate, audio handling, where each actually fits, and why WebRTC wins as the pragmatic default.
- The two voice loops. STT-LLM-TTS vs Speech-to-Speech, and what each means for the RTC layer.
- How the leaders build it. What OpenAI Realtime, LiveKit, Gemini Live, and ElevenLabs actually use at the edge, read from real behavior.
The levers
Tuning past the default
- Tuning past the default. Signaling without renegotiation, Opus configuration, silence and bitrate, RED and NACK, noise suppression, echo cancellation, and more.
- The metrics that matter. Naturalness and latency, turn-taking and barge-in, and what
getStatsgives you for free. - Live walkthroughs. A
chrome://webrtc-internalsdeep dive into LiveKit, Pipecat, and a DIY (OpenAI-style) stack, side by side.
Format
2 hours, live and interactive.
Private to your team.
Tailored emphasis. Tell us where your stack sits and we weight the session toward it.
Who delivers it
Philipp Hancke
Deep browser-level expertise in everything WebRTC. Co-author of the WebRTC courses, long-time contributor at the protocol and browser layer.
Tsahi Levent-Levi
Independent WebRTC analyst behind BlogGeek.me. Brings the industry context: who is doing what, where the market is moving, and what it means for your roadmap.
Request the workshop
Tell me about your team and where your stack sits. I will get back to you personally with scheduling and pricing.