Meta has introduced Muse, a personal AI agent designed to handle real-world tasks on a person’s behalf. Unlike simple chatbots, Muse actively completes work, manages schedules, and turns long-term goals into actionable plans. As a result, it aims to take everyday tasks off people’s plates entirely.

What Muse Does

Muse goes beyond answering questions. Instead, it takes direct action. For instance, it can send emails, book travel, and open a browser to fill out forms. Additionally, it can negotiate on a person’s behalf.

Once someone shares a goal, Muse helps build a personalized plan. Then, it coordinates time and resources before advancing the work independently. For longer tasks, Muse continues working even after the app is closed. It returns when something changes or when it needs approval, such as before sending an email or making a purchase.

Furthermore, Muse learns from conversations over time. It remembers details a person mentioned only once, so it can make unprompted suggestions later. For example, it can turn a saved recipe into a grocery list, suggest a dinner party menu, and factor in friends’ dietary restrictions before sending invitations.

How People Interact With Muse

Muse works the way people already communicate. Consequently, talking to it feels similar to messaging another person, whether through the Muse app or directly in WhatsApp. No technical experience is required to use it.

Behind the scenes, Muse is powered by Muse Spark, described as Meta’s most capable model built specifically for agentic, real-world work.

Payments and Purchases

When it’s time to check out, Muse can use Link, built by Stripe. Notably, it is the first AI agent covered by Link’s purchase protections. These include free coverage for damaged or lost items, price-drop refunds, no-fee returns, and a return guarantee on eligible purchases.

Link’s wallet for agents generates a one-time-use card, keeping real card details hidden during purchases. Meanwhile, Shop Pay is expected to arrive soon as an additional payment option, alongside 1Password support so Muse can use logins people already have.

Privacy and Security Built Into Muse

Since personal agents require a new kind of secure computing environment, Meta built Muse Secure VM specifically for this purpose. This dedicated cloud computer houses both the agent and a person’s data, isolated so no other agent can access it.

A separate system, called Sentinel, runs on the same machine but stays isolated from Muse at the system level. Nothing Muse does reaches the internet unless Sentinel approves it, and it asks for permission when needed.

Moreover, Muse has no visibility into passwords or payment methods. Any credentials a person shares go into secure storage, so Muse can use them without ever seeing them, even passwords typed directly into the browser.

Before sensitive actions like sending an email or making a purchase, Muse checks with the person first. It also provides a complete audit trail of everything it has done and plans to do.

People retain full control over the personal AI agent. They choose which apps Muse connects to and how much access it receives. For email specifically, users decide whether Muse can only read messages or also send them. Access can be changed or revoked at any time, and users can opt out of having their interactions used to train Meta’s AI models.

Additionally, Muse does not share conversations or VM data with Meta’s advertising systems. People can also tell Muse to forget specific things it has learned.

Later this year, Meta plans to introduce Muse Confidential VM. In this version, the entire VM, including data and conversations, will be encrypted with a key only the user holds, meaning even Meta will not be able to access it.

Availability

Muse is rolling out in the United States on iOS, Android, and muse.ai, with support for AI glasses coming soon. It is free for most everyday use, while subscription plans are available for people who want expanded capabilities.

Looking Ahead

Meta views personal superintelligence as one of the most transformative technologies of this era. With Muse, the company is taking an early step toward agents that handle more of the work, allowing people to focus on what matters most to them.

For more details on the design choices behind the system, Meta has published further explanations under “How We Built Safety Into Muse” and “How We Designed Muse.”

Overall, Muse represents Meta’s attempt to bring a personal AI agent into daily life, combining task automation, memory, and built-in privacy protections into a single assistant.

Related post

View all