Session abstract
What you’ll learn
The to-do list is one of the classic first projects for developers. In this hands-on workshop, we will make it talk back to us. Using JavaScript and Deepgram's Voice Agent API, you'll wire speech-to-text, an LLM, and text-to-speech together over a single WebSocket, then add items by voice, interrupt the agent mid-sentence and watch it yield, and (if time allows) give it a whole new personality. No voice AI experience is needed: the workshop starts from working code, every step is explained as you go, and everything works typed as well as spoken.