Meta Enables Voice and Video Calls with Muse Avatars

Meta has activated Muse Voice, a feature that lets U.S. users place real-time voice and video calls with Muse AI avatars that display animated facial expressions during conversation.

Muse Voice and underlying models

Muse Voice builds on Meta's Muse Realtime Avatar and Muse Realtime Voice models. The system listens to a user’s voice and drives a moving avatar that reflects conversational tone through facial expressions and subtle movements. Meta released the feature for users located in the United States.

Avatar personalization and design intent

The assistant appears as an animated avatar nicknamed Jolly, which users can personalize. Meta said Jolly's appearance can be configured as a human, an animal, or a fantasy character. Meta described the character-driven approach as intended to foster a deeper emotional bond between users and the technology.

Privacy, platforms and rollout plans

Meta said the models do not receive the user's image; they process the user’s voice to generate the avatar’s expressions. Alexander Wang, Meta's chief AI officer, announced the release in a post on X and on the company’s official blog.

• Real-time conversation support is enabled on mobile devices, the web, and macOS.
• Meta plans to extend Muse Voice to its Muse Charm wearable and to AI-enabled smart glasses before the end of the year.