Meta launches Muse personal AI agent with dedicated secure cloud VMs
Meta has introduced Muse, a personal AI agent designed to carry out tasks instead of only answering questions. Users can describe a goal through messaging, after which Muse may send emails, book travel, negotiate bills, complete forms, or work toward longer-term objectives. It continues operating after the app is closed and returns when user approval is required. Actions such as sending an email or completing a purchase are paused for confirmation, while an audit trail records what the agent has done and plans to do.
Each user receives a dedicated cloud virtual machine, called Muse Secure VM, where the agent, browser, and credentials are isolated. A separate Sentinel agent authorizes connector operations and network requests. Muse receives surrogate tokens rather than actual credentials, with real secrets injected at the network boundary. Its browser sub-agent uses an accessibility tree and cannot run JavaScript.
Muse is rolling out in the United States on iOS, Android, and muse.ai, with free and paid plans. Developers cannot self-host the consumer service, but Muse Spark 1.3 is available through Meta Model API and Muse Code. Meta says an open-weights release is planned.
Written by Sakhanda Wire AI from reporting by MarkTechPost. Facts, quotes and full context are in the original article below.