Meta unveiled Muse on September 8, a personal AI agent designed to complete real-world tasks rather than just answer questions. It can send emails, book travel, fill out forms, and coordinate multi-step goals across the apps people already use, continuing to work even after the app is closed.
The agent runs on a dedicated "Muse Secure VM" with its own isolated browser, and a separate "Sentinel" system monitors outbound actions and asks for approval before sensitive steps. Meta says Muse cannot see passwords or payment details, and that conversations are kept out of its ad-targeting systems. Payments go through Link by Stripe, with Shop Pay support coming later.
Muse is rolling out first in the US for users 18 and older, on iOS, Android and muse.ai, with AI glasses support planned. There's a free tier alongside $20 and $100 monthly plans for heavier use. Reviewers note Meta's security promises will need independent scrutiny, given the company's history of privacy settlements, including a $5 billion FTC penalty in 2019.