Meta has launched Muse in the United States as a personal AI agent designed to carry out tasks across connected apps and services rather than simply answer questions.
Each Muse runs inside a dedicated virtual machine with its own browser. Users choose which services the agent can access and can revoke those permissions. Meta presents that isolation model as a core security feature.
Reuters reports that Muse can handle workflows involving email, travel booking, payments and other app-based tasks. The launch follows additional security work after earlier internal testing reportedly exposed reliability and privacy concerns.
Muse is therefore a more consequential product surface than a standard chatbot: persistent cross-app access and the ability to act on a user’s behalf raise the value of the agent, but also the cost of permission mistakes or unreliable automation.