You can execute

session.generate_reply

before a participant joins. It does not require another participant to be present; the agent can use this method while alone in the room. This is a common pattern—for example, agents often call

generate_reply

in their

on_enter()

hooks or as proactive greetings and updates, regardless of participant presence. The method returns a

SpeechHandle

which you can manage as needed for the agent's conversational flow. This allows the agent to initiate conversation or provide updates even if no users have joined yet. Let me know if you need implementation details or examples of this pattern. Sources: examples/voice_agents/mcp/mcp-agent.py | External data and RAG | LiveKit Docs | Agent speech and audio | LiveKit Docs