ttsmodels.
M

MetaVoice-1B

Free

High-quality multilingual speech with emotional nuance.

MultilingualExpressiveOpen-Source

Overview

MetaVoice-1B is a 1.2B parameter model trained for emotional, natural speech with support for voice cloning and cross-lingual output. It targets American and British accents out of the box and can be adapted to others. The Apache-2.0 license makes it friendly for commercial use.

What makes MetaVoice-1B special?

  • Focus: Expressive multilingual speech.
  • Availability: Free & Open-Source (Apache-2.0).
  • Use cases: Voice cloning, dubbing, and expressive narration.
Visit official site

Related free models