THE ESSENTIALS IN 30 SECONDS
- Meta announced on November 5 that real-time voice and expressive avatars are rolling out in the Muse app, starting in the US.
- Users can choose voice only, an avatar beside the chat or full screen, pick from what Meta calls hundreds of voices, and interrupt replies.
- Meta says voice conversations are coming soon to its AI Glasses and Muse Charm, without giving a date.
Meta has started rolling out spoken conversations and animated avatars in Muse, the app for its AI assistant. The company announced the update on November 5, after previewing real-time Muse conversations at its Connect 2026 event. It is available on mobile, web and Mac, beginning in the US.
How the voice mode works
A conversation starts from a phone icon in the top-right corner of the app. Meta pitches the feature for thinking out loud, for example brainstorming, planning a trip or rehearsing an interview or presentation. Replies are spoken back immediately, and users can cut in while Muse is still talking. The company also says Muse can keep working on tasks in the background during a conversation, so requests can be made while walking or with busy hands.
There are three ways to view a session: audio only, an avatar shown next to the conversation, or the avatar in full screen. According to Meta, the avatar reacts while the user speaks and uses expressions and gestures matched to the tone of the exchange. Meta says there are hundreds of voices to choose from, in styles ranging from a New Yorker to a pirate, and the voice can be changed at any time.
The technology behind it
The feature relies on two systems Meta calls Muse Realtime Voice and Muse Realtime Avatar. A Meta research post from September describes the avatar model as an audio-driven diffusion transformer that generates video in short chunks from the voice model's stream of speech tokens, keeping lip movement and expression in sync with speech. Meta reports about 870 milliseconds between the end of a user's turn and the start of the response, with portrait video streamed at 25 frames per second. These figures, and the human-preference comparisons with competing avatar products in the same post, come from Meta's own evaluations. The post also says the Muse app is for users aged 18 and over and that generated video carries an invisible watermark.
What is still missing
Meta says voice conversations are "coming soon" to its AI Glasses and to Muse Charm, but gives no timeline. The announcement does not mention pricing or plans for countries beyond the US. iPhone in Canada notes that Muse does not appear on the Canadian App Store, even though the separate Meta AI app, which already has voice chat, is available there.
