To run code after a call has finished in a Python agent, you can use event handlers. The system typically uses an event-driven architecture where you can register handlers for specific events. For example: • To handle function calls completion, you can use the

@self._session.on("function_calls_finished")

decorator to register a handler that emits a

function_calls_finished

event with the list of called functions. • To handle response completion, you can use a handler like

_handle_response_done

to process response data and emit a

response_done

event when the response is completed. By registering your own event handlers or extending these existing ones, you can execute custom code after a call has finished. Sources: livekit.agents.multimodal.multimodal_agent API documentation | livekit.plugins.openai.realtime API documentation