Voice Agent Build Day

The next consumer interface is voice. Come build it. The frontier is realtime, full-duplex conversation: agents that listen while they speak, handle interruption gracefully, read tone and emotion, and respond in under a second. Voice quality is fast becoming table stakes. The hard part now is doing it at scale, for thousands of simultaneous users, without latency or cost falling apart. We're convening 100+ founders, engineers, and researchers in Hillsborough for a full day of building on that frontier. Builders get direct access to Inworld's Realtime API: hit Realtime TTS-2, speech-to-text, and a conversation-optimized LLM on their own, or run the entire speech-to-speech pipeline through a single API call. TTS-2 brings 100+ languages, natural-language voice steering, and roughly 200ms responses, with hands-on support from Inworld's engineering team. What you might build: a companion that notices you sound tired. A tutor that switches to Spanish mid-sentence without changing voices. A coach with an engaging personality. A support agent that's empathetic and adjusts to circumstance. A dungeon master that deepens the campaign experience. The common thread is realtime conversation. Space is limited. Apply to attend.