Meta’s new standalone AI app, Muse, hit the number 4 spot on the App Store within its first 24 hours.
The app launched on September 8, initially available only in the US across iOS, Android, and web at muse.ai. It’s not a chatbot. It’s not a search companion. Meta is calling it a “personal AI agent,” and the distinction matters because Muse is designed to actually do things on your behalf, not just talk about doing them.
What Muse actually does
The app is built to handle cross-service tasks that would normally require you to bounce between a dozen tabs. Booking travel, managing emails, filling out forms, making purchases.
Muse processes payments through Stripe Link integration, which means it can complete transactions without you manually entering card details every time. It also remembers user preferences over time, adopting a proactive approach to anticipating what you might need next.
The underlying technology comes from Meta’s Muse Spark family of models, developed by superintelligence labs led by Alexandr Wang. Previous releases of Muse Spark had already demonstrated the ability to boost app rankings significantly, so the rapid adoption pattern isn’t entirely surprising.
What sets Muse apart from the growing crowd of AI assistants is its architecture. Each session runs inside an isolated virtual machine, a sandboxed environment designed to keep user data compartmentalized. When the app encounters an action involving sensitive information, like a payment or a form with personal details, it pauses and asks for explicit confirmation before proceeding.
Meta has also been vocal about one particular design choice: Muse operates independently from the company’s advertising systems. Your interactions with the AI agent aren’t being funneled into the ad-targeting machine that powers Meta’s core business.
The pricing breakdown
Muse follows a freemium model with three tiers. The free version comes with weekly usage limits. The “Power” plan runs $20 per month for heavier usage. And the “Maximum” plan costs $100 per month for users who want essentially unlimited access to the agent’s capabilities.
Meta expects most users to stay on the free tier.
Why the timing matters
This launch doesn’t exist in a vacuum. Meta recently completed an approximately $18 billion multistate settlement related to social media usage issues. That settlement has clearly shaped how the company is positioning Muse from a trust and privacy standpoint.
The emphasis on isolated virtual machines, confirmation prompts for sensitive actions, and separation from advertising infrastructure reads like a company trying to get ahead of the inevitable regulatory scrutiny that comes with an AI agent handling your email, your purchases, and your personal information.
The competitive landscape is also shifting fast. OpenAI has been expanding its own agent capabilities. Meta entering with a standalone app, rather than embedding the agent inside Instagram or WhatsApp, is an interesting strategic bet. It forces users to make an active choice to download Muse rather than passively encountering it inside apps they already use.
Meta’s ability to promote Muse across its family of apps, which reach roughly 3 billion people globally, gives it a marketing funnel that most standalone app developers can only dream about.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
