Meta is rolling out real-time voice conversations and expressive avatars for its Muse AI assistant in the US, across mobile, web and Mac. Users can talk through an idea, interrupt Muse as it responds and choose from hundreds of voices, making this a more conversational interface than typing prompts into a box.
Meta AI Watch analysis
What happened
Meta says the new features are rolling out now in the US. To start a voice conversation in the Muse app, users tap the phone icon; they can choose voice only, show the avatar alongside the conversation or make it full screen. Meta says the avatar uses expressions and gestures that match the conversation, and users can change its voice whenever they like.
The company suggests uses including planning a trip, rehearsing a job interview or talking through a presentation. Voice conversations are coming to Meta’s AI Glasses and Muse Charm later, but Meta gives no date for that expansion.
Why it matters
Voice changes the shape of an AI interaction. Rather than composing a polished prompt, a user can start with a half-formed thought, hear a response and steer the conversation in real time. That could make Muse more useful for brainstorming or rehearsal, where a back-and-forth is the point rather than a tidy answer on screen.
The avatar and custom voices add another layer: this assistant is designed to feel present, not merely functional. That may make conversations more natural, but it also makes the design choices around an AI persona more consequential. The feature is rolling out in the US first, so this is not yet a universal Muse upgrade.
Our read
This is a meaningful change to how people can use Muse, not just a new coat of paint. Voice, interruption and flexible avatar display give users more ways to shape a conversation. Meta’s examples are plausible, but real usefulness will depend on how well the assistant follows a messy spoken train of thought, not how charming its pirate voice sounds.
What to watch
- Whether the US rollout expands to other regions.
- How well Muse handles interruptions and less structured spoken requests in everyday use.
- When voice conversations reach AI Glasses and Muse Charm.
Discussion spark: Do expressive voices and avatars make an AI assistant more useful, or risk making it feel more trustworthy and socially present than it really is?
Sources and evidence
- Source update (5 November 2026, 18:30 UTC)
not affiliated with, endorsed by, or operated by Meta