Voice AI

Voice agent demos are easy to build. A LiveKit worker, an STT → LLM → TTS pipeline, and you have something that talks. The hard part is making it reliable. We run AI interview agents on LiveKit Cloud. An interview lasts up to an hour. A lost recording is a compliance problem. A lost transcript means the candidate has to redo the interview. A silent agent means a human sitting in an empty room, wondering if anyone is there. ...

Voice agents are easy to demo and hard to scale. We landed on an architecture that keeps an agent natural, fast, and predictable: fire every possible response in parallel, let a fast router pick the winner. Here’s the journey, and why this design works. The problem nobody’s demo solves Every voice agent demo looks the same: clean audio, a cooperative user, a short scripted exchange. Then you ship it, and two things break immediately. ...