Particle.news

Meta Launches Muse, a Personal AI Agent That Can Act for Users

The U.S. rollout tests Meta’s Secure VM isolation, Sentinel gating and Stripe Link checkouts as the company tries to prove safety and win back user trust.

Overview

  • Meta released Muse in the United States on Tuesday, offering an agent that can open browsers, fill forms, send emails, book travel, negotiate on a user’s behalf and continue long-running tasks in the background.
  • Each Muse instance runs inside a dedicated cloud workspace called Muse Secure VM and is overseen by a separate Sentinel monitor that Meta says blocks unauthorized internet requests and prompts users before sensitive actions.
  • Muse can make purchases using Stripe’s Link, which issues one-time cards for agent checkouts and provides purchase protections, and Meta says credentials and payment details are stored so the agent cannot directly view passwords or real card numbers.
  • The product is free for light use but includes paid tiers reportedly priced at $20/month (Power) and $100/month (Maximum), and it is available via muse.ai, iOS, Android and WhatsApp with AI-glasses support and a Confidential VM promised later this year.
  • Journalists and internal posts report mixed lab results, including disconnects, task failures and at least one flagged security lapse, and those reliability concerns combine with Meta’s recent $18 billion multistate settlement to create a steep trust hurdle for consumer adoption.