You're asking whether an

agent.say

can be queued so it plays only after the user finishes speaking — yes. What you can do • Outgoing speech is automatically queued when the user is speaking if

allow_interruptions=True

(default). The agent will wait until the user's turn ends (VAD/STT or a custom turn detector) before playing queued speech. • To force immediate replacement of agent speech, call

session.interrupt()

. • To coordinate explicitly, listen for the user's turn end (e.g.,

UserStateChangedEvent

where

new_state == "listening"

) or use the

SpeechHandle

returned by

say()

to manage timing and playback. If you want, tell me which SDK/lang you're using (JavaScript, Python, or server-side) and I’ll provide a short code example. Sources: Turn detection and interruptions | LiveKit Docs | Agent speech and audio | LiveKit docs