Technology company Meta has announced the launch of Muse, a personal AI agent designed to proactively assist users with their goals, from routine tasks to long-term projects.
Unlike most existing AI assistants, Muse was built for a mass, non-technical audience: it requires no setup or technical experience and communicates with users the way people already talk to each other, either in a dedicated Muse app or directly within WhatsApp. The agent is powered by Muse Spark, Meta’s most capable model to date, which was developed specifically for real-world agentic work.
Users simply tell Muse what needs to be done. The agent can send emails, book travel, open browsers, fill out forms, and even negotiate on behalf of its owner. For larger ambitions, Muse helps develop a personalized plan, coordinates a user’s time and resources, and continues working autonomously — including after the app is closed.
It returns when circumstances change or when it needs approval, such as before sending an email or completing a purchase. Meta positions this as a way to achieve better outcomes with less effort, whether that means selling a car for a higher price, lowering a bill, or adjusting a training plan as a user’s schedule shifts.
Muse also maintains long-term memory of what matters to a person, allowing it to make unprompted suggestions and act on details mentioned only once — for example, turning a saved Instagram recipe reel into a grocery list or recalling guests’ dietary restrictions before sending dinner party invitations.
For payments, Muse integrates with Link by Stripe, which generates one-time-use virtual cards so real card details remain hidden. Meta notes that Muse is the first AI agent covered by Link’s purchase protections, including coverage for damaged or lost items and price-drop refunds. Support for Shop Pay and 1Password logins is coming soon.
Recognizing that personal agents require a new kind of secure computing environment, Meta built Muse Secure VM — a dedicated virtual machine in the cloud that houses both the agent and a user’s data, isolated so that no other agent can reach it. A separate Sentinel agent runs on the same machine at the system level, and nothing Muse does reaches the internet without the Sentinel’s approval.
Critically, Muse has no visibility into users’ passwords or payment methods; credentials are stored securely and can be used without being exposed. Users must confirm sensitive actions, receive a complete audit trail of the agent’s activity, and retain granular control over which apps Muse connects to and what it can do with each — access can be changed or revoked at any time.
On the privacy front, Meta states that Muse conversations and VM data are not shared with its advertising systems, and users can opt out of their interactions being used to train Meta’s AI models, as well as instruct Muse to forget specific things it has learned. Later this year, the company plans to introduce Muse Confidential VM, which will encrypt the entire virtual machine — including data and conversations — with a key held only by the user, meaning even Meta could not access it.
The new tool is rolling out in the United States on iOS, Android, and at muse.ai, with AI glasses support planned. It is free for most core functionality, with paid subscription tiers for heavier use. Meta frames Muse as a first step toward personal superintelligence, which it considers one of the most transformative technologies of the era — an agent that takes on more of the work so people can focus on what matters to them.
The post Meta Rolls Out Muse, A Proactive Personal AI Agent That Works Across The Apps People Already Use appeared first on Metaverse Post.
