I'm happy to help you get a text transcript of your speech-to-speech model interactions using LiveKit's built-in transcription features. Here's a simple guide to walk you through it: • Enable Transcription: First, make sure transcription is turned on in your

AgentSession

by setting

transcription_enabled=True

in

RoomOutputOptions

. This step is key to getting those transcripts. • Access Realtime Transcripts: You can then access the transcriptions as they stream over the

lk.transcription

text stream topic, which syncs nicely with audio playback by default. • Disable Synchronization (Optional): If you'd prefer to receive transcriptions immediately without waiting for sync, just set

sync_transcription=False

in

RoomOutputOptions

. • Retrieve Full History: To look back at the entire conversation, including all transcriptions, simply use

session.history

. • Realtime Monitoring: Stay updated by listening for events like

user_input_transcribed

and `conversation_item_added`—they let you know when new transcriptions come in. For a more detailed synchronization experience, you can enable

use_tts_aligned_transcript=True

in the Python SDK to get word-level timing with supported TTS providers. If you'd like, I’m here to provide more details or examples—just let me know! Sources: Docs::Agents | Session recording and transcripts | LiveKit Docs | Speech-to-text (STT) integrations | LiveKit Docs | main.py