LiveKit Community

use_tts_aligned_transcript=True

What does this do?

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


The TTS-aligned transcription feature improves the synchronization of transcription text with the agent's speech. It is particularly effective with TTS providers like Cartesia and ElevenLabs, which support word-level timing. When enabled, it ensures that transcriptions are displayed in sync with the audio output, enhancing the user experience.
 Sources: <https://docs.livekit.io/agents/build/text|Text and transcriptions | LiveKit Docs>

Why is it turned off by default? Does it cause any issues?

`use_tts_aligned_transcript` is turned off by default because it’s still an experimental feature that might change in future updates. Turning it on could cause some hiccups, especially if your TTS provider doesn’t support aligned transcripts or if the timing gets a bit off. You might also notice some formatting quirks or lose the original text’s style.
Right now, only a few TTS providers, like Cartesia and ElevenLabs, offer word-level transcription timing. For others, alignment happens at the sentence level, which can help with syncing but might not always be perfect.
If you’re thinking about enabling this feature, just double-check that your TTS provider supports aligned transcripts to keep things running smoothly.
 Sources: <https://github.com/livekit/agents/blob/3b9965981932666b52b262a54ec9955af11e5a09/livekit-plugins/livekit-plugins-cartesia/livekit/plugins/cartesia/tts.py#L1-L79|livekit-plugins/livekit-plugins-cartesia/livekit/plugins/cartesia/tts.py> | <https://docs.livekit.io/agents/build/text|Text and transcriptions | LiveKit Docs>