To help you fetch the complete audio of each turn in LiveKit, you can reconstruct it from individual audio frames. Let me walk you through how you can do this: • Use the

collect()

Method: The

ChunkedStream

class has a handy

collect()

method that gathers all audio frames and combines them into a single audio frame. This is probably the easiest way to get the complete audio output. • Reconstruct Audio from Events: If you’re working with integrations like AWS Bedrock, you can link incoming audio chunks to the active message by using the

audio_content_id

from the

audioOutput

event. Then, collect all frames associated with the same

response_id

and merge them with

rtc.combine_audio_frames()

. • Combine Frames: The

rtc.combine_audio_frames()

function is your friend here—it lets you merge a list of audio frames into one complete audio segment. This is also great for debugging or creating a full audio file. • Frame Collection System: You might find it useful to set up a system that listens to audio frame events during a turn, collects all frames, and then merges them into a complete audio segment. I hope this helps you reconstruct the complete audio for each turn smoothly. If you’d like me to clarify anything or need more guidance, just let me know—I’m here to help! Sources: livekit-plugins/livekit-plugins-google/livekit/plugins/google/beta/realtime...