It looks like the error 'property object has no attribute say' usually happens when the object you're trying to use

say()

on isn’t set up quite right or might be the wrong type. In the LiveKit framework, the

say()

method belongs to the

session

object, which needs to be properly initialized and started with audio output turned on. Here are a few steps that might help you fix this: • Make sure the

session

object is correctly initialized as part of the LiveKit framework. • Double-check that the session is started with audio output enabled by using

room_output_options=RoomOutputOptions(audio_enabled=True)

. • Confirm that the object you're calling

say()

on is actually a

session

object, and not something else like a string or an integer. If you’d like, feel free to share more about how you’re initializing and using the

session

object in your code—I’m here to help you work through it! Sources: Agent speech and audio | LiveKit Docs | complex-agents/role-playing/agents/narrator_agent.py | livekit-agents/livekit/agents/voice/agent_session.py