Voicebox Meta

Meta AI's generative speech model for in-context text-to-speech and style transfer

Voice & AudioFree

Voicebox is Meta AI's generative speech model trained on a large dataset of audiobooks in multiple languages. It performs in-context text-to-speech synthesis, noise removal, content editing, and style transfer without task-specific fine-tuning. As a research model, it represents a significant leap toward generalist AI voice systems that can match speaker styles across diverse samples.

Key Features

• ✓In-context TTS
• ✓Multi-language support
• ✓Style transfer
• ✓Noise removal
• ✓Content editing