Meta introduced Muse, a personal AI agent for people 18 and over, at its annual Connect conference on September 23, 2026 — and CEO Mark Zuckerberg called it the “centerpiece” of the company's AI roadmap, according to the Associated Press.
Unlike a chatbot that answers questions and waits for the next prompt, an agent is built to act on a person's behalf: opening a browser, filling out forms, booking a reservation, or working through a multi-step goal without being walked through it line by line.
What Muse Actually Does
Muse is available now in the U.S. through a dedicated app or inside WhatsApp. Per Meta's own announcement (cited by the Associated Press), the agent can handle small requests — sending an email, booking travel — as well as longer-running goals, such as building a year-long exercise plan or helping set up a new business. Meta has partnered with services including OpenTable, Spotify, and Ticketmaster so the agent can complete real bookings and tasks rather than just suggest them.
Meta is also stressing the security side of the launch: Muse reportedly runs on a dedicated, secure virtual machine that houses both the agent and the user's data, and Meta has been explicit that the agent is restricted to users 18 and older for now.
Beyond the App: Glasses and a Standalone Device
Meta said it is working to bring Muse to its AI glasses lineup for hands-free use, and Zuckerberg announced a separate palm-sized device called Muse Charm, arriving in time for the December 2026 holidays, that packs the assistant into a keychain-sized gadget usable without a phone, computer, or glasses.
The launch sits inside Zuckerberg's broader public vision, laid out in a lengthy essay published on Meta's site in August 2026, of a “personal agent” that works continuously on a user's behalf across relationships, career, finances, and daily logistics. It's a notably different posture than rivals like Anthropic, which has publicly called for a more coordinated, cautious approach to advancing AI capability — Zuckerberg used the Connect keynote to distance Meta from that camp directly.
Why This Matters Beyond Meta
Muse is part of a wider shift the whole industry is making: from AI that answers questions to AI that completes tasks. That's the same direction voice and chat-based AI agents for business are heading — systems that don't just respond to a customer, but actually complete a booking, qualify a lead, or resolve a support ticket end-to-end. Consumer launches like Muse tend to speed up that expectation: once people get used to an agent handling errands in their personal life, they expect the same from the businesses they deal with.
For companies evaluating whether to add an AI agent to their own website, WhatsApp line, or call center, the practical questions are usually less about the underlying model and more about integration: does the agent actually complete the task (book the appointment, answer the billing question) or does it just chat convincingly and hand off to a human anyway? That gap is exactly where a well-built agent earns its cost.
Frequently Asked Questions
Is Muse available outside the United States?
Not yet. As of its September 2026 launch, Muse is only available to users in the U.S., per Meta's own announcement.
Do I need Meta AI glasses to use Muse?
No. Muse currently works through a standalone app or inside WhatsApp; glasses support and the standalone Muse Charm device are planned as additional ways to access it, not requirements.
Is Muse free to use?
Meta has not published separate pricing for Muse access itself at launch; some related hardware, like the Ray-Ban Meta Audio glasses and hearing-enhancement feature announced at the same event, do carry their own separate costs.
How is this different from a regular AI chatbot?
A chatbot answers a message. An agent, like Muse, is designed to carry out multi-step tasks on its own after being given a goal — for example, planning and booking a trip rather than just describing travel options.
