I’ve been trying a few setups, I’m curious: what the lowest e2e turn latency you got? what was the setup (self-hosted vs livekit cloud, choice of models, etc.)?
What I mean by turn latency is the time between you stop talking to the moment you first hear the answer of the bot.