Thanks — you’re asking how to make an agent deliver a greeting. The simplest, most reliable way is to generate the greeting when the agent becomes active by overriding the agent’s on_enter lifecycle method and calling session.generate_reply(). Why use on_enter: • It runs as soon as the agent session starts, so it’s ideal for initial greetings. • It uses the agent’s LLM to produce a natural, context-aware message (better than a static playback if you want conversational tone). Quick example (Python-style):

async def on_enter(self) -> None:
    """Called when the agent session starts. Generate initial greeting."""
    await self.session.generate_reply(
        instructions="Hello! I'm your assistant. How can I help you today?",
        allow_interruptions=False
    )

Notes and alternatives: • If you only need a one-off greeting outside an Agent class (for example in your session entrypoint), you can call

await session.generate_reply(...)

after

session.start(...)

. • Use

allow_interruptions=False

if you want the greeting to finish before the user can interrupt. • If you need non-LLM static audio or pre-recorded prompts, play that audio in the same on_enter method instead of using

generate_reply()

. Would you like a short example for your specific stack (JavaScript/TypeScript or Python) or help with making the greeting multilingual or retry on no-answer? Sources: Build a multilingual voice agent that automatically switches languages | Li... | Agents and handoffs | LiveKit Documentation | Agent speech and audio | LiveKit Documentation | How to implement call greeting retries in voice agents | LiveKit | Agents telephony integration | LiveKit Documentation