NotebookLM conversational speech model
Text-to-speech (TTS) with Next-gen Kaldi
Generate spoken audio from text using selectable voices