Meta rolled out Muse Voice, a video chat feature for its consumer AI agent Muse, to all U.S. users on Thursday. The feature, first shown at Meta's Connect developer conference in September, pairs new AI models called Muse Realtime Avatar and Muse Realtime Voice technology to power an animated digital avatar that users can customize and interact with through live conversations.
The avatar, nicknamed Jolly, displays subtle expressions and gestures that match the tone of conversation, though it does not see the user. Meta positions the visual avatar as a way to deepen user connection to the AI agent compared to voice-only competitors. Users can change the avatar's appearance to resemble a person, animal, or other character of their choice.
Alexandr Wang, Meta's chief AI officer, announced the rollout via blog post and X, sharing a demo where his avatar appeared as a fluffy white dog. In the clip, Wang joked about his meme-posting habits on X and his Muse suggested adding disclaimers to avoid misinterpretation.
The feature works on mobile, web, and Mac, with future support planned for Meta's AI Glasses and an upcoming wearable device called Muse Charm, which Meta plans to launch before year-end. The Charm device will display the avatar on a small screen for voice and task-based interactions.