Meta Will Bring Its Muse AI Agent to AI Glasses

Meta is bringing Muse, the AI agent that can book travel, make phone calls, send emails and complete purchases on a user's behalf, to its AI glasses in "the coming months." The company announced the expansion at its Connect 2026 event on Wednesday, extending an agent that launched earlier this month from standalone iOS, Android and Mac apps, and WhatsApp chats, onto the face-worn hardware at the center of its AI hardware strategy.

Muse for glasses will let users delegate tasks entirely by voice. Saying the agent's name and making a request activates it, after which it works in the background and checks back in when the task is done. Meta framed the glasses version of the agent as a way to "stay on top of things, take tasks and projects off your plate and turn long-term goals into action plans, all hands-free."

The stakes here are straightforward. Agentic AI — software that takes multi-step actions across services rather than just answering questions — is the current battleground for consumer AI, and the major players are racing to prove their agents can be trusted with real-world errands. Meta's move pairs that ambition with its most controversial hardware category. The pitch is that users could have Muse book a flight while jogging, with no phone in hand.

What Muse can do

Muse's task list is broad. According to Meta, the agent can use the web, book travel, make phone calls, send emails and even make purchases on a user's behalf — provided the user is willing to trust it with the most sensitive parts of their digital life. In the standalone apps and WhatsApp, that functionality is already live. The glasses version adds a hands-free interface to the same set of capabilities.

At Connect, Meta offered concrete examples of what the glasses integration will handle: booking appointments, checking flight prices and tracking down items. Meta also suggested use cases that lean on the glasses' always-worn form factor, including guiding personalized workouts, logging meals and providing nutritional insights.

Permission and control

Meta has built consent mechanisms into Muse's design. The agent asks for permission before carrying out tasks that involve sensitive actions, such as making purchases. Users choose which outside services Muse can access, and they can revoke those connections at any time.

Those controls matter because an agent that can spend money, send communications and act across third-party services concentrates an unusual amount of personal authority in one piece of software. Meta's revocation option is the main lever users have to pull Muse's access back once granted.

The convenience question

It is not obvious that glasses add much to what Muse already does. The core appeal of agentic AI is that it works in the background without constant supervision. Pulling out a phone to give an agent a new task is not a heavy burden, and Muse's existing apps already offer the full agent experience on those devices.

The glasses make the interaction hands-free, and Meta's suggested scenarios — workouts, meal logging, nutritional guidance — are cases where a phone is genuinely inconvenient. But those read more as conveniences than necessities. The value proposition narrows to situations where your hands are occupied or your phone is put away.

Hardware and privacy

The deployment also extends Meta's agent deeper into hardware that has already drawn criticism. The glasses have attracted a privacy backlash — detractors have called them "perv glasses" — precisely because they carry cameras and sensors into social settings. Layering an agent with access to email, phone calls, purchases and outside services on top of that hardware raises the same questions with higher stakes: the device now not only sees the world around its user but can act on the user's behalf across their digital accounts.

Meta's answer is the permission system — opt-in access to outside services, confirmation before sensitive actions, revocable connections. Whether users find that sufficient will shape how quickly agent-on-face becomes normal rather than novel.

Why it matters

The announcement signals where Meta believes consumer AI agents are heading: off the phone screen and onto the body. If voice-activated, background-running agents prove useful on glasses, Meta gains a differentiated surface for its AI that rivals without comparable hardware cannot easily match. If the hands-free agent remains a marginal convenience, the glasses become a smaller footnote in the agent race.

For now, Muse for glasses exists as a promise with a loose timeline. "The coming months" is the only schedule Meta has offered, and the company has not said whether every Muse capability — purchases, calls, email — will be available on the glasses from day one. Users who already own Meta's AI glasses, and are comfortable with the privacy tradeoffs, will effectively get the company's new personal agent bundled into the experience. The rest of the market will be watching whether hands-free delegation becomes a habit or stays a demo.