To configure a LiveKit agent to only respond when specifically called upon, follow these steps: 1. Set Up Explicit Agent Dispatch: Configure the agent with an explicit dispatch rule by setting the

agent_name

in the

WorkerOptions

. This disables automatic dispatch. Example:

if __name__ == "__main__":
    cli.run_app(WorkerOptions(
        entrypoint_fnc=entrypoint,
        agent_name="my-specific-agent"  # Disables automatic dispatch
    ))

2. Implement Conditional Response Logic: Modify the agent's instructions to only respond to specific triggers. Example:

class MyAgent(Agent):
    def __init__(self):
        super().__init__(
            instructions="You are a helpful agent. Only respond when explicitly called.",
            stt=deepgram.STT(),
            llm=openai.LLM(model="gpt-4o-mini"),
            tts=cartesia.TTS(),
            vad=silero.VAD.load(),
        )
        self.waiting_for_trigger = True

    async def on_enter(self):
        self.session.say("Waiting for trigger...")

3. Set Up Request Handling: Use the

request_fnc

in

WorkerOptions

to control when the agent accepts or rejects job requests. Example:

async def handle_request(request: JobRequest):
    if request.metadata.get("trigger_condition"):
        await request.accept(
            name="my-specific-agent",
            identity="my-specific-agent-id",
        )
    else:
        await request.reject()

if __name__ == "__main__":
    cli.run_app(WorkerOptions(
        entrypoint_fnc=entrypoint,
        request_fnc=handle_request
    ))

This setup ensures the agent only responds when explicitly called. Let me know if you need further clarification or additional examples. Sources: Anatomy of an Agent | LiveKit Docs | Agents telephony integration | LiveKit Docs | examples/voice_agents/toggle_io.py