Meta rolls out real-time voice conversations for Muse
Meta announced a voice and expressive-avatar rollout for Muse on November 5, 2026, letting users have real-time voice conversations with the agent. Meta says the feature is rolling out to Muse users in the US.

Key takeaways · 4
- 01
The rollout is for Muse users in the US; Meta announced it on November 5, 2026.
- 02
Real-time conversations are supported on mobile, the web and Mac.
- 03
Muse can continue working on tasks in the background during a conversation.
- 04
During voice chats, Muse hears users but does not see them.
What Muse’s voice mode includes
Meta CEO Mark Zuckerberg announced that Muse users can have real-time voice conversations with the agent.[2] Meta says Muse Voice lets users call and video chat with its consumer AI agent, while the mode also offers an animated avatar.[3] Users can choose voice-only interaction, show the avatar beside the conversation or switch to a fullscreen view.[1] Meta says the avatar’s expressions and gestures match the conversation’s tone.[3] The company had previewed the voice and avatar experience at Connect on September 23, 2026.[1]
Availability and how to start
Meta says the feature is rolling out to Muse users in the US, with real-time conversations supported on mobile, the web and Mac.[2][3] A phone icon in the top-right corner starts a conversation.[1] Meta says voice conversations are coming soon to its AI Glasses and Muse Charm device.[3] The Verge reports that Muse Charm is supposed to ship this year.[2] The evidence does not give a price or a more specific timetable for the rollout or the upcoming device support.
What users can do—and what Muse sees
Meta presents voice mode as a way to brainstorm ideas, plan trips and rehearse interviews, presentations or pitches.[1] The agent can continue working on tasks in the background while the conversation continues.[1] During voice chats, Muse does not see users, but it hears them and responds based on the conversation.[3] Meta says users can select from hundreds of voices and change their selection.[1] That distinction matters when deciding whether the mode suits a task that depends on visual input: the evidence describes voice interaction, not Muse seeing the user.
How Meta describes the technology
Meta says Muse Realtime Voice and Muse Realtime Avatar power the new mode.[4] The company says sharing the token stream keeps voice, lip motion and expression synchronized.[5] Meta reports that the avatar generates portrait video at 448 by 768 pixels and 25 frames per second, with about 870 milliseconds of response latency.[5] It also says its process reduces model evaluations per chunk from 120 to two, and that its optimizations increase serving capacity eightfold over a two-step BF16 baseline.[5] Meta says avatar videos receive a durable, invisible watermark.[5] In a brief test, The Verge’s Jay Peters said Muse responded quickly, though it sometimes paused while retrieving information or checking something.[2]
Voice and an animated avatar give professionals another interaction format to evaluate for brainstorming, planning and rehearsal. Teams should weigh that option against the stated limitation that Muse does not see users during voice chats, and against the reported pauses when it retrieves information.
Why it matters
Bring this to your team — one session a day, built for your industry.
kju.ai is built for teams: daily sessions, team analytics, and rollout support for your organization.
For teamsHow this developed
6 November 2026
Meta rolls out real-time voice conversations for Muse
Sources
- Meta Brings Real-Time Voice and Expressive Avatars to Muse – Unite.AIunite.ai
- Muse has a voice mode now | The Vergetheverge.com
- Meta says US users can now video chat with their Muse AI avatar | TechCrunchtechcrunch.com
- Meta rolls out Muse voice and expressive avatars in the UStestingcatalog.com
- Bringing Your Muse to Life | Meta AI Researchresearch.meta.ai