Audio credits

Lesson audio is synthesized from the public lesson text. These are synthetic voices, not original speaker recordings.

• English: Kokoro, British English voice bf_emma. Kokoro model (Apache 2.0); Kokoro ONNX.
• Spanish: Davefx; German: Thorsten; Slovak: Lili. The voice model cards describe the respective datasets as CC0.
• French: SIWIS voice. SIWIS French speech synthesis database, licensed under CC BY 4.0. The speech model is adapted from this dataset.
• Arabic: Kareem. Dataset and training information: Arabic TTS training project.

The non-English model files and individual model cards are provided by Piper voices. Synthesis uses Piper (GPL 3.0 or later). Alphabet pronunciation overrides are configured separately from ordinary text synthesis.