Audio & software credits

• Pronunciation clips are computer-generated, not recordings of a human tutor.
• English: Piper en_US-ljspeech-high, trained on the LJ Speech public-domain dataset.
• Japanese, Korean, Mandarin: MeloTTS (MIT License) — github.com/myshell-ai/MeloTTS
• Hindi: Kokoro-82M (Apache License 2.0; hexgrad) — huggingface.co/hexgrad/Kokoro-82M
• Missing clips use the browser’s built-in Speak / Web Speech voices.
• License notices: MIT · Apache 2.0

Clearance for public hosting reviewed 2026-09-25 (portfolio / education risk posture).