Meta Launches Muse, an AI Agent That Sends Emails and Books Travel

Meta launched its personal AI agent Muse in the United States on Tuesday, designed to handle everyday tasks on a user's behalf: sending emails, booking travel, filling out forms, and negotiating bills. It's the company's most ambitious step yet toward Mark Zuckerberg's vision of "personal superintelligence."

Muse connects to apps and services across several categories at once: email, calendar, payments, health apps, shopping, and smart home devices. Users decide which apps the agent can access and can revoke that access at any time.

What the Agent Can Do

Beyond simple tasks like sending an email or booking a flight, Muse is meant to carry out longer-term goals, such as building a year-long workout plan or helping set up a new business. The system opens a browser, fills out forms, negotiates contract or bill terms on the user's behalf, and, thanks to its integration with Link by Stripe, can also make payments.

Among its more everyday functions, Meta cites turning a recipe from a Reels video into a shopping list or sending out party invitations. The company says Muse will eventually come to Meta's AR glasses too, moving interaction with the agent beyond the phone screen.

Safety Under Scrutiny

Every Muse agent runs on its own isolated cloud virtual machine, letting it keep working on tasks in the background even when the user isn't actively using the app. Meta says the system has no access to the user's passwords or payment methods and doesn't share data with the company's advertising systems.

Those assurances have met with skepticism from commentators, who point to Meta's track record on privacy: a 2011 settlement with the US Federal Trade Commission over misleading consumers, a record $5 billion fine in 2019 for privacy violations, and the Cambridge Analytica scandal. TechCrunch summed up that context succinctly, invoking an industry refrain about the gap between the company's statements and its practices.

Everyone will have an exceptionally capable personal agent that understands you, your goals, and everything you care about. Your agent will work 24/7 on your behalf to improve your relationships, health, career, finances, home management, hobbies, and more - Mark Zuckerberg, CEO of Meta

The Race for Personal Agents

Muse enters a market where major industry players are competing ever more fiercely for the role of everyday, ever-present AI assistant. Its rivals include Google's Gemini Spark and Anthropic's Claude Cowork, along with a growing number of web browsers built with AI agents and integrations with messaging apps like iMessage and SMS.

The project is built on the Muse Spark family of models, which Meta's head of AI, Alexandr Wang, has been rolling out at a rapid pace since April 2026. The agent fits into Zuckerberg's broader vision, laid out in a 6,500-word essay published in August 2026, in which the Meta CEO predicted an era in which everyone would have the tools to start a business, get PhD-level tutoring, and receive personalized life advice.

What It Means for Users

For now, Muse is available only in the United States, so Polish users don't yet have direct access to the service. Still, the scope of the permissions the agent receives over email accounts, calendars, and payments raises a question that will sooner or later reach Europe too: how to reconcile the convenience of autonomous AI agents with personal data protection under the GDPR and the EU's AI Act.

Muse's rollout also shows that the race between Meta, Google, and Anthropic is shifting from chatbots that answer questions to agents that actually act on a user's behalf in the real world, with access to their inbox and their wallet. It's another step toward a model in which an AI assistant doesn't just advise but makes decisions and carries out transactions without ongoing human oversight.