The tutor's voice

When the tutor speaks, the audio is generated on our own server by open-source models, so what you hear isn't sent to a speech company. Depending on the server it uses Kokoro-82M (Apache-2.0) or Piper with the LibriTTS-R medium voice, trained on LibriTTS-R by Koizumi et al., licensed CC BY 4.0. Voices are synthetic and can mispronounce names or terms.

· Artelos">
Artelos
Sign in Try the demo