Muse is 📝Meta's personal AI agent, launched in September 2026, that pursues a user's tasks and goals autonomously through a messaging interface, running on a dedicated cloud virtual machine.
Muse is a consumer-facing personal 🏷️#ai-agent designed to require no technical setup. Users talk to it as they would message a person, in the Muse app or inside 📝WhatsApp. It is powered by Muse Spark, which Meta describes as its most capable model to date. Muse can send email, book travel, open a browser, fill out forms, and negotiate on a user's behalf; for larger goals it drafts a personalized plan and continues working after the app is closed, returning when something changes or approval is needed.
Each user's agent runs in its own Muse Secure VM, which stores the agent, user data, and connected-service credentials. A separate Sentinel agent on the same machine must approve any outbound internet action. Muse cannot see passwords or payment methods, asks before sensitive actions such as sending email or purchasing, and keeps an audit trail of past and planned actions. Meta states that VM data is not shared with its ad systems and that users can opt out of model training; a Muse Confidential VM with user-held encryption keys is planned for later in 2026.
Payments run through @Link by 📝Stripe, which issues one-time-use cards and extends purchase protections to Muse, the first agent covered. Shop Pay and 📝1Password support are announced as coming soon. Muse also remembers user details for unprompted suggestions, such as turning a saved 📝Instagram recipe reel into a grocery list. It is rolling out in the US on iOS, Android, and muse.ai, free for most use with paid tiers, and is slated for AI glasses.