Muse – Meta’s personal AI agent
Points and comments are a snapshot, not live.
Meta released Muse, a personal AI agent that browses the web and takes actions for you.
Muse is a personal AI agent from Meta that runs on a dedicated virtual machine with its own browser, capable of booking appointments and handling customer service. It works across text messaging and WhatsApp, requiring user approval for sensitive actions like purchases or sending emails. User logins are stored in a credential store the agent cannot read; 1Password integration is planned. Muse can be connected to email, calendar, Instagram, and other apps. A free tier exists with usage limits; a paid subscription is available.
What commenters are saying
Dominant reaction: strong distrust of Meta's data practices. Many commenters argue they would not trust Meta with the level of access a personal AI agent requires. Several note that despite privacy claims, Meta has a long history of bad faith behavior. A few defend the product on technical grounds: Meta benefits from an enormous existing user base of 3 billion people. Some commenters who tried Muse say it aggressively pushes purchasing actions, one reported it prompted to set up buying functionality immediately. A camp of technical users compares Muse to other agentic tools like OpenClaw, Grokbot, and Instinct, noting Meta's advantage in distribution.