TTSLabs

Listen to a document or webpage instead of reading it on screen.

Generate expressive, multilingual speech from a short voice reference with an open local model.

Generate natural speech locally or in a public demo with a small open-weight TTS model.