Text-to-Speech (TTS) for Min Nan / Taiwanese Hokkien using Meta’s MMS-TTS Model (facebook/mms-tts-nan)

🗣️ This application uses Meta's multilingual speech model (MMS-TTS) to generate natural speech from text written in Min Nan Chinese (zh-nan), also known as Taiwanese Hokkien. Input text should be written using the Pe̍h-ōe-jī (POJ) romanization system.

Example input: Lí hó! Góa sī lâng Tâi-oân.

The output is a 16kHz WAV audio file synthesized using the VITS-based neural TTS model.