Meta Lets Users Call and Video Chat With Its Muse AI Assistant in Real Time

Meta has begun rolling out real-time voice conversations and video chat capabilities for its Muse artificial intelligence assistant, allowing users to interact with an animated AI avatar instead of relying exclusively on text messages. Announced on November 5, 2026, the new feature introduces Muse Voice, which lets users speak naturally with the assistant, receive spoken responses and interrupt conversations whenever they want. The rollout initially covers users in the United States through the Muse application on mobile devices, the web and Mac.

The update represents another step in Meta’s ambition to make AI assistants more interactive and accessible through everyday devices. Powered by its Muse Realtime Voice and Muse Realtime Avatar technologies, the system combines conversational AI with an animated character that responds using facial expressions and gestures. Users can customise the avatar’s appearance and voice, choosing from hundreds of voice options. Meta also plans to bring voice conversations to its AI glasses and upcoming Muse Charm wearable device.

Meta Introduces Real-Time Voice and Video Conversations With Muse

Until now, users primarily interacted with Muse through written instructions and responses. The latest update adds a more conversational experience.

Users can tap the phone icon in the top-right corner of the Muse application to start a voice conversation.

Once the call begins, they can speak naturally rather than preparing detailed written prompts. Muse listens, responds in real time and can continue working on tasks while the conversation is taking place.

For example, someone planning a holiday could explain their preferences verbally, discuss destinations and ask the assistant to develop a travel plan. Similarly, users can rehearse job interviews, practise presentations or brainstorm ideas through an interactive conversation.

One important capability is interruption support. Users do not have to wait until Muse finishes speaking before asking another question or changing the direction of a conversation.

This makes the experience closer to speaking with another person than interacting with a traditional text-based chatbot.

Muse Gets an Animated Avatar That Responds to Users

The most distinctive part of the update is the animated avatar accompanying Muse’s voice.

Unlike conventional AI voice assistants that respond through audio alone, Meta’s system gives users a visual character that moves and reacts during conversations.

The avatar displays facial expressions and gestures designed to match the tone of the conversation. It also reacts while users are speaking, providing visual feedback that the assistant is following the discussion.

Meta initially introduced an animated character called Jolly, but users can customise their Muse avatar to look like a person, an animal or another character.

The company offers three ways to experience the conversations:

• Voice only: Speak with Muse without displaying the animated character.
• Avatar alongside chat: Keep the character visible next to the conversation.
• Full-screen avatar: Display the animated character across the screen for a more face-to-face experience.

The customisation options are intended to make interactions feel more personal and give users control over how their AI assistant appears.

Users Can Choose From Hundreds of AI Voices

Meta is also introducing extensive voice customisation.

Users can choose from hundreds of voices, allowing them to select a personality and speaking style that suits their preferences.

The company has demonstrated different voice characteristics, including regional styles and playful personalities. Users can change their selection whenever they want.

This is important because voice is becoming a major part of the competition between AI assistants.

Companies are increasingly moving beyond text-based interfaces towards systems that can communicate through natural speech, respond quickly and maintain conversational context.

Meta’s approach combines voice selection with a persistent animated avatar, creating a more personalised experience than a conventional voice assistant.

The Technology Behind Muse Realtime Voice and Avatar

The new experience is powered by two technologies developed by Meta Superintelligence Labs: Muse Realtime Voice and Muse Realtime Avatar.

Muse Realtime Voice handles the conversational intelligence and generates speech for the interaction.

Muse Realtime Avatar converts the speech and conversational signals into corresponding visual movements, including facial expressions and gestures.

According to Meta’s technical explanation, both systems share a synchronised stream of speech information. This helps align the avatar’s movements with its spoken responses.

The company says the system can generate portrait-format video at 25 frames per second, with approximately 870 milliseconds of latency between the end of a user’s turn and the first response.

Meta has also described infrastructure optimisations designed to support multiple simultaneous conversations while keeping response times low.

The technical approach is important because real-time video generation requires considerably more computing resources than producing a text response or audio alone.

Muse Is More Than a Voice Assistant

The latest update builds on Meta’s broader vision for Muse as a personal AI agent.

The company introduced Muse in September 2026, describing it as an assistant capable of completing tasks rather than simply answering questions.

Muse is designed to work across connected applications and services, helping users organise projects, manage activities and carry out online tasks.

Its capabilities include:

• Planning trips and organising travel activities.
• Helping users manage emails and other digital tasks.
• Supporting online shopping and product discovery.
• Organising information and creating action plans.
• Working on longer assignments in the background.

Meta says Muse operates through a dedicated secure virtual machine with its own browser. Users control which applications it can access and what permissions it receives.

The company also says the assistant requests approval before sensitive actions, such as sending emails or making purchases.

The addition of voice and animated avatars creates another way to interact with these capabilities without having to type instructions.