Meta Launches Personal AI Agent Muse, Free to Start with Monthly Plans Up to $100
nashnova research
Meta launched Muse, a personal AI agent that sends emails, books trips, and checks out purchases on your behalf — free to start, up to $100 a month at the top tier. The catch: it demands deeper data access than any consumer social product before it.
What can Muse actually do?
The core idea is "do it for you," not "chat with you." Muse connects to your email, calendar, payments, smart home, and other apps, then executes tasks directly.
Use cases include sending emails, booking travel, lowering bills, filling out forms, and even turning recipe videos into shopping lists — then checking out via Stripe's Link.
Shopify Shop Pay and 1Password integrations are coming soon. This means → Muse is not a chatbot; it is a digital butler that operates your apps on your behalf.
How does pricing work — will most people pay?
Free to start, but you must add a payment card. Two paid tiers: Power at $20/month, Maximum at $100/month — the difference is task-processing capacity.
Meta says most users will stay on the free tier. The app includes a usage meter and warns you before any charge kicks in.
In plain terms = get users in free, charge when they use more — the same playbook as cloud computing's pay-as-you-go model.
Why does it need so many permissions?
Muse asks for access to email, calendar, payments, health and fitness, and more. Users can authorize apps one by one. If a service has a public API, Muse connects via user credentials; if not, it accesses through a browser.
This means → you are not just granting read access to data. You are letting an AI agent log in and act as you.
This is the deepest permission request of any consumer AI product to date — the trust bar is extremely high.
Can the privacy promises be trusted?
Meta says Muse runs inside a dedicated "Muse Secure VM," with a separate Sentinel Agent enforcing system-level isolation. Muse itself cannot access passwords or payment details, and user data will not feed into Meta's ad system.
These technical claims have not yet been independently verified. The bigger issue is Meta's track record — the company has faced years of data-misuse controversies, and Muse demands a level of trust far beyond the social-media era.
This reflects a fundamental tension: the company best positioned to scale an AI agent is the one users trust least with their data.
Where do the competitors stand?
The space already has several players: Google's Gemini Spark can operate web pages, Anthropic's Claude Cowork targets collaborative tasks, and various AI agents are embedded in iMessage, WhatsApp, and other chat platforms.
Meta's differentiator is sheer user scale — Facebook, Instagram, and WhatsApp reach billions, and Muse will also come to Meta AI glasses.
In plain terms = the technology gap between rivals is narrow. Meta is betting that "users are already here, switching costs are low." Whether it can clear the trust barrier is the make-or-break test for Muse.
市场有风险，内容仅供研究参考，不构成投资建议。