Meta launches Muse AI agent to automate digital tasks and transactions
The new 'agentic' AI moves beyond chat to execute complex goals using a secure virtual machine and Stripe integration.
Meta has launched Muse, a personal AI agent designed to shift the user experience from conversational AI to autonomous task execution. The system is currently rolling out in the U.S. across iOS, Android, and the web, with plans to integrate the technology into AI glasses in the future.
Powered by the "Muse Spark" model—which Meta describes as its most capable model specifically built for agentic workloads—Muse can manage complex goals, coordinate schedules, and handle online transactions. To facilitate financial autonomy, Meta integrated Link by Stripe, which provides purchase protections and one-time-use cards to keep a user's actual credit card details hidden during transactions. As Meta noted in its official announcement, the tool "doesn’t just answer questions, it actually does the work."
Security and Infrastructure
To address the security risks inherent in autonomous agents, Muse operates within a "Muse Secure VM." This dedicated virtual machine isolates the agent and user data from other agents on the system. Additionally, Meta implemented a "Sentinel" agent that serves as a system-level gatekeeper. The Sentinel agent is responsible for approving all internet requests and prompting the user for explicit permission before the AI performs any sensitive actions.
Beyond the cloud-based agent, Meta has also released "Muse Glimmer," a 30-billion-parameter open-weight model designed to enable local AI agents to run on consumer hardware.
The Strategic Pivot
The launch of Muse represents a broader strategic pivot by Meta to compete with rivals like Google and OpenAI. The company is moving away from simple large language model (LLM) chatbots toward "agentic" AI—systems capable of performing tangible work, such as filling out forms, booking travel, or negotiating bills. This shift is supported by massive infrastructure investments and a push toward what the company views as "personal superintelligence.
This transition is further evidenced by Meta's acquisition of the AI agent startup Manus in December 2025, a deal reported to be valued between $2 billion and $3 billion.
Industry Implications
If successful, Muse transforms Meta's AI from a feature embedded within social apps into a central operating layer for a user's entire digital life. By solving the primary hurdles of trust and security through secure VMs and the Sentinel gatekeeper, Meta is attempting to normalize the use of AI for autonomous financial and personal management.
What's Next
Industry observers are now watching for the rollout of Muse into wearable hardware, specifically AI glasses, which would move the agent from a screen to a real-time environmental interface. While the core functionality is now live in the U.S., the scale of adoption will depend on whether users trust the Sentinel system to manage their sensitive data and finances autonomously.