Meta's personal AI agent, now with a real-time video-chat avatar and customizable voice — a generated, expressive face you speak to, with synced lips, interruptible replies, and voice-only, side-by-side, or full-screen modes, live for US users on mobile, web, and Mac.
Last verified · 2026-11-06 · by Moe Ameen
Meta Muse is the personal AI agent from Meta Superintelligence Labs. On November 5, 2026, Meta rolled out Muse Voice — real-time voice plus a generated, video-chat avatar — to all US users, after previewing it at Meta Connect in September 2026. Two new models drive it: Muse Realtime Voice and Muse Realtime Avatar, which share a single stream of speech tokens so the spoken reply and the avatar's video stay in lockstep — lips, expressions, and gestures matched to what's being said.
In practice you tap a phone icon to start talking, can interrupt the agent mid-answer with a follow-up, and Muse can keep working a background task while the conversation continues. Three display modes let you choose how present the face is: voice only, the avatar beside the chat, or the avatar full screen. The avatar is generated by the model rather than driven by your camera, and you can set its appearance — human, animal, or another character — and pick from a large set of voices, swapping the voice at any time. It's available on mobile, the web, and Mac for users 18 and older; Meta says versions for its AI glasses and the Muse Charm wearable are coming later.
Notably, this is a conversational avatar. It's a face you talk to, and an interface for the agent — not a presenter that fronts finished, published video for an audience. Meta watermarks the generated avatar video via Meta Video Seal. Some precise specs floated around the launch (exact resolution, frame rate, latency) come from single reports, so treat those particular figures as provisional.
That boundary is the important part for a creator: Muse's avatar and voice make talking to the agent more natural, but they generate no branded media, hold no brand-voice system, and publish nothing to your channels. The content operation around it is still yours to run — which is where a content engine picks up.
The fastest front door to a piece of content is often your own voice, and that's exactly what Muse's new mode is good at — so use it as a capture surface, not a production one. Talk to the avatar the way you'd talk to a smart colleague: spill the idea, argue the angle out loud, let it push back and tighten the thread into an outline or a caption draft. What you walk away with is raw material. What Muse can't do is turn that into anything your audience sees: it renders no [Persona Short](/glossary/persona-shorts), composites no avatar into a branded template, lays out no [Carousel](/glossary/hyperframes), and pushes nothing to a single platform. Its avatar is pointed at you; a creator's avatar needs to point at the feed.
That's the clean hand-off to [Kompozy](/). Take the outline or draft from your Muse conversation into Kompozy as the seed for a content unit, and it fans into the formats a chat avatar can't make: captioned [Persona Shorts](/glossary/persona-shorts) and [Persona Frames](/glossary/persona-frames) — HeyGen avatar video with one face kept consistent across every clip and composited into brand-exact HyperFrames — alongside Carousels, Quote Graphics, Photo Posts, per-platform text, a Blog Article, and an Email Newsletter, all held to your voice by a [Persona Brief](/glossary/persona-brief). Then [Autopilot](/glossary/autopilot) schedules and publishes the set across the eight social platforms plus blog and email from one queue behind a per-post review. The through-line is simple: Muse turns your voice into a draft; Kompozy turns that draft into an on-brand, published avatar that actually fronts your content. For the agent's other features, see [Meta Muse](/ai-tools/meta-muse) and the [AI-agents early-access wave](/ai-tools/meta-muse-ai-agents).
It's a real-time, generated video-chat avatar for Meta's Muse agent, launched broadly to US users on November 5, 2026. Powered by Muse Realtime Voice and Muse Realtime Avatar, it gives the agent an expressive, customizable talking face you converse with — with synced lips, interruptible replies, and voice-only, side-by-side, or full-screen modes.
No. The Muse avatar is a conversational interface — a face you talk to — not a presenter that fronts published videos for an audience. It generates no posts, carousels, or scheduled content. To put a consistent on-camera persona into finished Reels and Shorts, you use a content engine like Kompozy, which generates Persona Shorts and Persona Frames and publishes across platforms.
Treat it as a voice-capture front door. Talk through an idea with the avatar to produce an outline or draft, then hand that seed to Kompozy, which generates the avatar video, carousels, images, blog, and newsletter and schedules them across platforms. Muse captures the thinking; Kompozy produces and ships the content.
At launch it's live for US users 18 and older on mobile, the web, and Mac; Meta says AI glasses and the Muse Charm wearable come later. It's part of the Muse agent, which Meta says is free for most needs with paid plans reported at $20 and $100 per month. Confirm current details on Meta's pages.