Is it ok that on the frontend "lk.transcription_final" is always "false" for user transcriptions?

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

Is there a way to get segment id of a transcription on the agent's side? I mean the segment ids we get within lk.transcription on the frontend.

I'm happy to help you access the segment ID of a transcription on the agent's side! You can do this by retrieving the

lk.segment_id

attribute from the text stream's attributes. Here's a simple example in Python using the LiveKit Agents SDK to guide you through it:

from livekit.agents.types import ATTRIBUTE_TRANSCRIPTION_SEGMENT_ID

async def on_text_received(reader: rtc.TextStreamReader, participant_identity: str):
    segment_id = reader.info.attributes.get(ATTRIBUTE_TRANSCRIPTION_SEGMENT_ID, None)
    if segment_id:
        # Handle the segment ID
        pass

This segment ID is unique for each transcription segment and is great for correlating updates. If you have any questions or need a hand with anything else, please feel free to ask—I’m here to support you! Sources: examples/other/chat-stream-receiver.py | VoiceAgent/Chat/Receive/TranscriptionStreamReceiver.swift | Transcriptions | LiveKit Docs | livekit/livekit_models.pb.go

I wasn't able to receive any events from "lk.transcription" on agent's side, but it works on the frontend.

Hi there! Just to clarify, the agent doesn’t directly "receive"

lk.transcription

events the way the frontend does. Instead, the agent sends transcription data through text streams, and the frontend or client needs to actively subscribe to the

lk.transcription

text stream topic to get those updates. If you’re not seeing these events on the agent’s side, here are a few things to check: • Make sure you’re not using outdated events like

RoomEvent.TranscriptionReceived

. Switching to text streams is the way to go. • Confirm that the frontend or client is subscribed to the

lk.transcription

topic. • If audio output is turned off, you’ll want to set up a custom receiver to listen to the

lk.transcription

text stream topic. • On the agent side, keep an eye on the

conversation_item_added

event to know when text input or output gets added to the chat history. Please feel free to reach out if you’d like a hand setting up the right subscription or if you want to troubleshoot this further—I’m here to help! Sources: How to Set Up Speech-to-Text Input with Text-Only Output in LiveKit Agents | Agents 0.x migration guide | LiveKit Docs | Agents v0.x migration guide - Node.js | LiveKit Docs | Text and transcriptions | LiveKit Docs | Text and transcriptions | LiveKit Docs

I mean this example https://github.com/livekit/agents/blob/f2fb5705c865ee7f220f60b5baeaf4c716db4cf6/examples/other/chat-stream-receiver.py#L1-L73 examples/other/chat-stream-receiver.py is not working for me.

Hi there! It looks like the

chat-stream-receiver.py

example might be having some trouble, and that can happen for a few reasons. Let’s walk through some things to check together: • First, please make sure the

LIVEKIT_URL

environment variable is set up correctly. • When you run the script, don’t forget to provide a room name as a command-line argument. • Also, double-check that the sender is publishing text to the right topic, which should be

lk.transcription

for agent output. • It’s important that the sender includes necessary attributes like

segment_id

in the text stream. • And if you’re testing with the playground UI, just a heads-up: it doesn’t support text-only agent responses sent to

lk.transcription

. If you’d like, I’m here to help you verify your setup or troubleshoot any other issues you might be facing. Just let me know! Sources: examples/voice_agents/README.md | How to Set Up Speech-to-Text Input with Text-Only Output in LiveKit Agents