Gradium Open-Sources Phonon, a 31M-Parameter On-Device TTS Model Under Apache 2.0

Gradium has open-sourced Phonon, a compact on-device text-to-speech model with 31M parameters under Apache 2.0, delivering natural English speech without cloud dependency or GPU. Phonon runs efficiently on CPUs, phones, Raspberry Pi, browsers, and NVIDIA GPUs, achieving competitive accuracy and fast audio generation.