Meta launched Muse on Tuesday, a personal AI agent that connects to a user's email, calendar, payments, and other services to handle everyday tasks autonomously. The launch comes less than two weeks after Meta agreed to an $18 billion multistate settlement over social media's consumer harms, and the product demands significantly more trust from users than social media ever did.
What Muse Actually Connects To
Muse works by linking to the apps and services people already use. The agent has built-in connectors for common services and can set up connections to anything with a public API using credentials the user provides. When no API exists, it accesses the service through a browser. Users connect apps one at a time, making the opt-in nature visible rather than buried in settings. The agent can send emails, book travel, fill out forms, turn recipe reels into grocery lists, send invitations, and make purchases through Stripe's Link. Shopify's Shop Pay and 1Password support are on the roadmap.
The agent continues working after the user closes the app, returning when something changes or when it needs approval for actions like purchases. Over time it learns from conversations to make unprompted suggestions based on what matters to the user.
The Security Architecture
Meta built Muse on a dedicated virtual machine called Muse Secure VM, isolated from other users. A separate agent called Sentinel runs on the same machine but is kept apart at the system level, monitoring everything Muse does and blocking any action that would reach the internet without explicit approval. Muse has no visibility into passwords or payment methods, and Meta claims it does not share conversations or data with its advertising systems. A confidential version of the VM, encrypted so that even Meta cannot access it, is planned for later this year.
The opt-in design extends to training data. Users can opt out of having their interactions used to train Meta's AI models and can instruct the agent to forget specific information it has learned about them.
Pricing and Availability
Muse is available in the US on iOS, Android, and muse.ai, with WhatsApp integration and Meta AI glasses support coming soon. The base tier is free, with two paid plans at launch: Power at $20 per month and Maximum at $100 per month. Both offer increased usage for task handoff. Meta expects most users to stay on the free tier, and the app includes a usage meter that warns when free limits are approaching. A payment card is required to get started regardless of plan.
The Trust Problem
Muse enters a market where several competitors offer similar agent functionality. Google has Gemini Spark, Anthropic has Claude Cowork, and Microsoft has Copilot Tasks. The differentiator Meta is pushing is consumer accessibility, but the trust barrier is real. Meta settled with the FTC in 2019 over privacy violations for $5 billion, was charged in 2023 with violating that same privacy order, and discovered user passwords stored in readable formats in 2019. The Cambridge Analytica scandal still shapes public perception. For an agent that needs access to email, payments, and personal data to function, those history lessons matter. Meta is betting that the utility of having an AI handle daily tasks, combined with the ability to customize the agent's name and appearance, will be enough to overcome that skepticism. Whether it will is an open question.