It would be great to integrate EVI with an avatar and have realistic lipsync. However, to do this we would need to have the same Word or Phoneme timestamps that are provided in your TTS functionality - https://dev.hume.ai/docs/text-to-speech-tts/timestamps