AI voice and avatars

Meta Muse now has a voice: real-time calls with an avatar that talks back

Meta Muse voice went live on November 5, 2026, turning Meta's AI agent into something you can phone: you can call it, pick from hundreds of voices, interrupt it mid-answer and watch it reply through a live animated avatar. For now it is US-only, but the technology behind it, Muse Realtime Voice and Muse Realtime Avatar, says a lot about where AI avatars and clones are heading.

Published on 5 min readBy the editorial team of The AI Creator System

Leggi questo articolo in italiano

AI-generated cover: a stylized face of light in profile, releasing cyan and violet sound waves on a midnight blue background, with the title Meta Muse now talks
Cover image generated with artificial intelligence. How we use AI

What Meta announced on November 5, 2026

Meta Muse now talks: on November 5, 2026 Meta rolled out real-time voice conversations and expressive avatars in the Muse app, two features first shown at Connect 2026. The official announcement, titled "Give Your Muse a Voice", is on Meta's newsroom.

Muse is Meta's AI agent: not just a chatbot, but an assistant that can also carry out tasks. Voice changes how you use it: instead of typing, you call it. To start, you tap the phone icon in the top-right corner of the conversation.

The technology has two names: Muse Realtime Voice, which generates the voice, and Muse Realtime Avatar, which generates the animated face. We covered the model behind Muse in our article on Meta Muse Wonder.

How Muse voice works

Muse voice works like a phone call, not like a voice note. According to Meta's announcement, you can:

  • choose from hundreds of voices and styles, such as a New Yorker accent or a pirate, and switch whenever you like;
  • interrupt Muse while it is answering, the way you would with a person;
  • let Muse work in the background while you keep talking, for example on a walk.

Meta says voice conversations are coming "soon" to AI Glasses and Muse Charm, with no date. The first release covers the Muse app on mobile, web and Mac.

The launch is US-only. As PPC Land reports, Muse is available in Canada but voice is not active there yet. Meta has not announced dates for Europe or other countries.

The real-time avatar: what is behind it

The avatar is optional and comes in three modes: voice only, avatar next to the chat or full screen, for something closer to a video call. According to Meta, gestures and expressions follow the tone of the conversation, and the avatar reacts while you are the one speaking.

The technical details are in the research post "Bringing Your Muse to Life", published by Meta on September 23, 2026. The voice produces a stream of speech tokens that carry both the words and how they are said; the avatar reads the same stream, which keeps lips, expressions and voice in sync. Meta's stated figures are these:

MetricStated value
Response timeabout 870 milliseconds after you finish speaking
Video format448×768 pixels, portrait, 25 frames per second
Sessions per GPU12 real-time avatars on a single NVIDIA GB200
WatermarkMeta Video Seal, invisible, applied to the generated video

Meta also compared its avatar with two competing products in 2-to-3-minute conversations: in its tests, raters preferred Muse over Runway Characters 78% of the time and over HeyGen LiveAvatar 88% of the time. These are internal evaluations, not verified by third parties; on mannerisms, the gap with Runway is not statistically distinguishable from parity. The same post notes that not every example shown matches the avatars available in the app today, and that Muse is intended for users aged 18 and over.

What the announcement leaves out

Meta's announcement leaves some important questions open. It gives no price or usage limits, does not list supported languages and does not explain how voice recordings are handled: how long they are kept and whether they are used to train models. PPC Land also points out that it is unclear whether voice conversations stay out of advertising systems.

There is no API either: Meta's research does not describe a way to use Muse Realtime Avatar in your own apps. Today it is a feature of the Muse app, not a tool for developers or creators.

What it means for you as a content creator

The message for creators is clear: AI avatars are moving from recorded video to live conversation. When Meta, the company behind Facebook, Instagram and WhatsApp, puts an animated face and a natural voice inside a mass-market app, audiences quickly get used to talking to a digital face. If you already have your own avatar or clone, you start ahead.

Muse, however, does not give you your own face and voice: the avatars belong to Meta and the app only works in the US for now. If you want a clone of yourself for your own videos, you build it today with the tools already available: the Create your AI clone in 30 minutes course walks you through it step by step, and our AI clone guide compares the options. If you would rather not do it yourself, there is our 30-second clone avatar service, done for you.

On the voice side, also read our articles on ElevenLabs' Eleven Realtime and on ChatGPT's new voices with GPT Live 1.1. If you want a complete method that brings together avatars, video, voice and selling, there is The AI Creator System.

Frequently asked questions

When did Meta Muse get voice?

Meta announced real-time voice conversations and expressive avatars for Muse on November 5, 2026, after previewing them at Connect 2026.

Can I use Muse voice outside the US?

Not yet. The launch is US-only, on mobile, web and Mac. In Canada Muse is available but voice is not active, and Meta has not announced dates for other countries.

How many voices does Meta Muse have?

Meta says there are hundreds of voices and styles, which you can switch at any time.

Do I have to use the Muse avatar?

No. You can talk with voice only, show the avatar next to the chat or go full screen.

Can I use the Muse avatar for my own videos or to build my clone?

No: the avatars are those of the Muse app, and Meta has not released an API for Muse Realtime Avatar. For a clone with your own face and voice you need other tools.

Read next

Sources checked on November 8, 2026