Flux TTS and Pipecat were both designed for how real conversations actually work. Deepgram's TTS product lead Jeff Liu and Daily's CEO Kwindla Kramer show you what the stack makes possible, live.

Date: Tuesday, September 29, 2026
Time: 10:00 AM PT | 1:00 PM ET
Duration: 60 minutes
Format: Live Technical Webinar

Jeff Liu, Staff Product Manager, TTS, Deepgram
Kwindla Kramer, CEO, Daily

Deepgram built Flux TTS for the conversation itself: expressive by default, consistent across turns, and native to interruption and turn lifecycle. Pipecat, the open orchestration framework from Daily, gives developers the pipeline to wire voice agents together in real time. Bring the two together and you get a stack where the voice and the framework were both designed for how real conversations actually work.

This session is a tour of what that stack makes possible. We'll walk through the kinds of agents you can build with Flux TTS as the voice and Pipecat as the framework, show a live demo of one running end to end, and unpack what's happening under the hood: how a persistent connection carries tone across turns and keeps latency consistent from the first response to the fiftieth, how interruption feedback tells your agent exactly what the caller heard when they barged in so it picks up intelligently instead of losing its place, and how the pipeline handles the entity-heavy, multi-turn moments that trip up most voice stacks.

Join Jeff Liu, Deepgram's TTS product lead and Kwindla Kramer from the Pipecat team at Daily for a live, hands-on look at what the two can do together.

• See a real agent in action: Live demo of a production-style voice agent built with Deepgram and Pipecat, tested against interruptions, entity-heavy responses, and long multi-turn calls.
• What Flux TTS brings to a Pipecat pipeline: A conversation-native TTS changes how your agent handles interruptions, sounds across a long call, and stays responsive turn after turn.
• How to wire it up: Swap Flux TTS into an existing Pipecat pipeline in a few lines using the official service integration, with Pipecat orchestrating turn lifecycle and interruption events over Flux's conversation-native protocol.
• Code you can build on: Concrete examples and reference code you can lift straight into your own Pipecat pipeline.

This session is built for developers and engineering leads shipping real-time voice agents, especially teams already using Pipecat or evaluating it. If you're wrestling with TTS that flattens across turns, breaks on barge-ins, or needs constant tuning to sound human, this is for you.

Save your spot to see Flux TTS and Pipecat in action, and walk away with reference code you can build on the same day.