Meta Opens Muse Avatar Video Calls to All US Users, Preps Wearable Device
Starting Thursday, US users can talk by voice and video with Meta's AI agent, which now features real-time facial expressions and will arrive on smart glasses and the Muse Charm device by year's end
Meta on Thursday (5) opened unrestricted access to Muse Voice for all users in the United States. The feature enables voice calls and video calls with the company's artificial intelligence agent, which now displays an animated avatar with facial expressions and gestures synchronized to the tone of the conversation.
The announcement was made by Meta's AI director, Alexandr Wang, in a post on the company's official blog and in a demonstration on X — in which his avatar, a fluffy white dog, suggested he add warnings to the memes he publishes.
Two models behind the conversation
The experience is powered by two proprietary models unveiled at the Connect conference in September: Muse Realtime Avatar, responsible for the character's facial and body animation, and Muse Realtime Voice, which processes speech in real time.
According to Meta, the avatar — nicknamed Jolly — does not "see" the user. It only hears and reacts to audio, modulating expressions and gestures according to the tone of the conversation. Personalization is unrestricted: users can turn the digital companion into a person, an animal, or any other character.
The company's bet is that anthropomorphizing the agent this way will increase the consumer's emotional bond with AI — a connection likely to intensify as interactions shift from text to live conversations.
Availability and next steps
Real-time conversations work on mobile apps, on the web, and on Mac. Meta plans to bring the feature to its AI Glasses and to the "Muse Charm," a wearable with a small Tamagotchi-style screen that the company intends to launch before the end of the year.
When the hardware arrives, the digital avatar will appear on the device's screen so the user can converse with the agent, which will be able to respond or perform tasks on their behalf.
The conversational agent race
The move positions Muse in the contest for voice and video interfaces with AI assistants, but with a clear differentiation: instead of a disembodied voice, Meta bets on a customizable visual character as an engagement layer.
The strategy also lays the groundwork for the company's hardware ecosystem, which seeks to transform the software agent into a physical presence in users' daily lives — from glasses to the pocket wearable.
The expansion to all US users comes weeks after a limited initial launch and signals that Meta considers its real-time inference infrastructure mature enough for national scale.