Meta is putting its personal AI agent in a place that makes the idea feel much more consequential: in front of your eyes, throughout the day.
At Connect 2026, the company said Muse will come to its AI glasses in the coming months. The move extends Muse beyond phones and computers, allowing users to call up the agent hands-free and ask it to act on what they are looking at. A product on a shelf, a flyer on a wall or a long school-supply list could become the starting point for a task rather than something the user has to photograph, describe and process manually.
That is a meaningful change in interface. The camera and microphone are no longer just ways to capture content or issue commands. They become the sensory layer through which an agent understands a person’s immediate surroundings.
Muse is moving from chatbot to background assistant
Meta says Muse will support longer conversations while working in the background. It is also adding connectors for services including Walmart, Best Buy, Sephora, Expedia, Instacart, Notion, GitHub and Box, alongside payment support from Shop Pay and PayPal.
The details matter because they show what Meta means by a personal agent. Muse is not being positioned only as a voice assistant that answers questions. It is being built to move between shopping, travel, household tasks and work while the conversation continues. Meta will also give Muse its own email address, creating another route for the agent to receive instructions or complete work.
That ambition depends on a growing hardware family. Meta says it will offer more than 100 AI-glasses styles across Ray-Ban, Oakley and Meta Glasses by the end of the year. The lineup now includes Ray-Ban Meta Audio glasses, Ray-Ban Meta Gen 3 and new collaborations, while Meta VR Glasses are being positioned as a private cinema, workspace and gaming device.
The real product is access, not eyewear
The broader shift is easy to miss if the announcement is read as another glasses launch. Meta is trying to make AI available without requiring a user to stop, open an app or frame a carefully written prompt. The goal is ambient access: the agent is available because the device is already being worn.
That could make AI more useful in small, ordinary moments. It could also make the boundary between asking for help and delegating action much thinner. Meta’s earlier Muse announcement says users control connected apps and permissions, while sensitive actions such as sending email or making a purchase require confirmation. Those controls will matter more when the agent can interpret the world around someone and act without a phone screen in the loop.
Meta’s strategic bet is clear: the winning AI interface may not be a better chatbot window. It may be a device that stays present until the user needs it.