Meta launches Muse personal AI agents, with free tier and $20 or $100 plans

Meta introduced Muse, a personal AI agent app that can handle tasks like booking appointments, completing electronic forms, and monitoring home security camera feeds. The app, available to U.S. consumers, will launch with a free tier and paid options priced at $20 or $100 per month based on usage.

Muse runs in an isolated environment

Meta says Muse operates within an “isolated environment” inside its computing infrastructure. The company states the agent “never sees your actual passwords or payment details” and prompts users before performing sensitive actions.

Meta also requires users to opt out of letting it use their interactions to train its AI models. If users do not opt out, Meta will remove “critical personally identifying information” from conversations and interactions before using them to improve its models, according to David Singleton, Meta vice president of engineering.

Pricing, availability, and what the agent can do

Muse is powered by the Muse Spark family of foundation models. Meta gave the project the internal code name Hatch.

The app will be available on iOS, Android, and via a standalone website for U.S. consumers, with later support through Meta’s Ray-Ban Meta glasses. Meta also plans a WhatsApp version, which Wang said will share the same core conversation functionality but will not include the feed or an ideas tool that recommends AI agent “trips and tricks.”

Users can view a feed with updates aggregated from connected Facebook and Instagram accounts, or summaries of certain articles or stories from the web depending on what they want the agent to collect.

Security review and bug bounty

Meta is asking third-party security researchers to test Muse for vulnerabilities through a bug-bounty program with financial rewards. The company said it has hardened Muse through “dogfooding,” agentic red teaming, and issues found in real adversarial scenarios through its private bug bounty program.

Launch amid legal and scrutiny pressure

Meta’s launch arrives as the company faces heightened scrutiny across AI and platform safety. Meta recently agreed to pay nearly $17 billion in a settlement with state attorneys general over allegations that it misrepresented the prevalence of harm on apps like Facebook and Instagram, and it still faces other personal injury and school district lawsuits tied to similar claims.

Wall Street pressure also remains a factor as Meta works to diversify beyond digital ads and show returns on AI infrastructure spending. Meta CEO Mark Zuckerberg has previously said “new personal agents” will support the company’s next wave of products and revenue, and the company is exploring monetization models such as taking a cut of AI agent-related shopping transactions, though no plan has been finalized.

Why it matters

Muse is Meta’s push into personal AI agents that can take actions across everyday tasks, while the company tries to manage security, privacy, and regulatory risk at launch. The pricing structure and the opt-in approach to training data will shape how quickly users adopt the product and how regulators and security researchers evaluate it.