Voicebox

by Meta AI · open-source · Last verified 2026-03-17

Voicebox is Meta AI's generative speech model based on non-autoregressive flow matching that achieves state-of-the-art performance on text-to-speech, noise removal, content editing, and style transfer tasks through a unified in-context learning approach. Its flow-matching architecture allows it to generalize to new voices and styles without fine-tuning, setting a new paradigm for zero-shot speech synthesis.