how can I know if the speech has been spoken to the user or not?

Thanks for reaching out, bogdanmelnikbo. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.

What is the proper way to use it in python? Is there example?

Hey there! Let me walk you through tracking speech playback in Python using LiveKit's awesome

SpeechHandle

class. I've got some super handy examples that'll make this super clear. • *Basic awaitable pattern:*`handle = session.say("Hello world") await handle # Waits for speech to finish playing print("Speech playback complete")` • *Status checking:*`if handle.done(): print("Speech has finished playing") if handle.interrupted: print("Speech was interrupted")` • *Using a callback:*`def on_speech_complete(h): print(f"Speech {h.id} done") handle = session.say("Processing your request...") handle.add_done_callback(on_speech_complete)` • *Advanced: ensure speech is played before moving on*`# Wait for the current speech to complete before ending a call current_speech = session.current_speech if current_speech: await current_speech.wait_for_playout() print("Agent finished speaking before call ended")` These patterns are super helpful – they'll let you know exactly what's happening with your speech playback, so you can smoothly handle different scenarios. Pretty cool, right? Source: official LiveKit documentation Sources: Agent speech and audio | LiveKit Docs | livekit-agents/livekit/agents/voice/events.py

in that case how to understand that speech has been started? Just started, not finished

Hey there! Want to know how to catch the moment when your agent starts speaking? I've got two super handy methods that'll help you track speech playback in Python. 😊 First up, you can listen for the

speech_created

event on your

AgentSession

. It's like a little notification that pops up right when the agent begins talking:

from livekit.agents import AgentSession, SpeechCreatedEvent

@session.on("speech_created")
def on_speech_created(ev: SpeechCreatedEvent):
    print("Agent started speaking")
    # ev.speech_handle provides the speech handle
    # ev.user_initiated tells if it was user-initiated
    # ev.source shows the origin (e.g., 'generate_reply')

Not feeling that method? No worries! You can also track

agent_state_changed

events and check when the state switches to `speaking`:

from livekit.agents import AgentStateChangedEvent

@session.on("agent_state_changed")
def on_agent_state_changed(ev: AgentStateChangedEvent):
    if ev.new_state == "speaking":
        print("Agent started speaking")

Either way, you'll get an instant heads-up when speech playback kicks off. Pretty cool, right? 🎉 Source: official LiveKit documentation Sources: Turn detection and interruptions | LiveKit Docs | livekit-plugins/livekit-plugins-azure/livekit/plugins/azure/stt.py