Meta Introduces Muse: an AI Agent That Actually Gets Things Done

Meta is taking its vision for personal AI a step further with Muse, an agent designed to handle tasks instead of simply answering questions. It can plan goals, browse the web, manage connected services, make purchases, and remember useful details while giving users control over sensitive actions. Behind it all is a security system built to protect personal data, credentials, and conversations as Muse works in the background.

Meta announced that Muse  is already rolling out in the US on iOS, Android, and muse.ai, and coming soon to AI glasses. It’s free for most of what people need, with subscription plans for people who want to do more.

Muse is designed to handle both everyday tasks and larger goals with minimal input. Users can ask it to send emails, book travel, open websites, fill out forms, negotiate on their behalf, or coordinate plans around a broader objective. It can continue working after the app is closed and return when it needs approval or when something changes. Muse can also remember useful preferences and details, allowing it to make proactive suggestions, such as turning a saved Instagram recipe into a shopping list or considering friends’ dietary restrictions when planning a dinner. For purchases, Muse can use Link by Stripe, with a one-time-use card that keeps users’ real payment details hidden, while Shop Pay and 1Password support are planned.

Privacy and security are central to how Muse operates. Meta built Muse Secure VM as a dedicated virtual machine where the AI agent, user data, and connected service credentials are isolated from other agents. A separate Sentinel agent controls Muse’s access to the internet and can require user approval before sensitive actions are completed. Muse cannot directly see passwords or payment details stored in secure storage, while users can decide which apps it connects to and what permissions it receives. The system also provides an audit trail of actions, allows users to disconnect services or change permissions, and gives them the option to prevent their interactions from being used to train Meta’s AI models. Meta says Muse conversations and VM data are not shared with its advertising systems, and users can ask Muse to forget specific information.

Meta sees Muse as an early step toward what it calls personal superintelligence, with the goal of having AI take on more of the work involved in everyday life. The company is also developing Muse Confidential VM, which will encrypt the entire virtual machine with a key controlled by the user, meaning even Meta would not be able to access the contents. Muse is initially rolling out in the United States on iOS, Android, and the web at muse.ai, with support for AI glasses planned. The service will be free for most common uses, while subscription plans will be available for people who need greater capabilities.