Meta's Muse Can Now Talk With You Out Loud, and Act for You

Meta now lets people hold real-time voice conversations with Muse, its personal AI assistant, Mark Zuckerberg announced.
The Verge reported that the voice mode and animated avatars are rolling out now in the US, inside the Muse app. While the assistant speaks, its avatar moves in step with the words.
You can change the assistant's voice tone by asking for a different one, or by using the options in the avatar settings.
The avatar runs on a system called Muse Realtime Avatar. Meta Superintelligence Labs, the company's AI research group, calls it state-of-the-art embodiment technology, which means technology that gives an AI a face or body. It is linked to Muse Realtime Voice, and Meta says it turns that voice into expressive, interactive avatars.
Voice has been developing for several months. In June, Meta introduced Meta AI Voice Conversations, powered by Muse Spark. Users could interrupt the assistant, change topics, or switch languages in the middle of a conversation.
In October, Zuckerberg's Meta Connect event included voice chat for Muse. Meta's AI voice mode also arrived on Messenger, Facebook, WhatsApp and Instagram DMs, where the assistant answers out loud. Meta says it is faster and more conversational, and it works in 13 languages. A setting in the Meta AI app called "Ready to talk" turns voice on by default.
Muse goes further than a chat assistant. Reuters reported that Muse can book trips, make purchases, schedule appointments and make phone calls. In September, Reuters reported that it could also reach other apps to send emails and make payments. It launched in the US through a dedicated app and on WhatsApp. Meta introduced Muse Spark earlier this year as its most powerful model so far.
Meta is also putting the assistant into hardware. It says Muse voice mode and expressive avatars are coming soon to its AI glasses. They will also be available on Muse Charm, a handheld AI device that Meta expects to ship this year. Zuckerberg's October post said Muse Charm would launch by December.
In my view, the bigger change is the mix of speech and action, not the avatar. A voice assistant that only answers questions is like a well-read friend. One that can also book the table, pay the bill and call the restaurant is something else. Its mistakes stop being wrong answers and start being wrong actions. Voice also removes the screen, where you would normally check a confirmation before anything happens. How Muse handles permission, interruptions during a payment, and recovery from errors may matter more to people than the animated face.
The broader context here is a shift from typing to ongoing spoken conversations, with a steady face for the assistant. For developers and businesses, the practical questions are how quickly speech can respond, how much access the agent gets to other apps, and whether this voice and avatar technology will be offered beyond Meta's own apps and devices. For everyday users, the appeal is simple. You can speak in your own language, interrupt when you need to, and change the tone without digging through menus. That lowers a barrier that has kept many people from using these tools at all. The benefit is real. So is the responsibility that comes with an assistant that can spend money and contact people on your behalf.


