Meta’s Muse Takes AI to the Next Level as Your Personal Assistant

Meta made a statement it hasn't made in a long time. After launching on 8th September, the company's new personal AI agent Muse has already achieved more than 900,000 downloads in its first week, which made it the No.1 App on the iOS App Store and Google Play in the US. The number of downloads grew to between 2.5 and 2.8 million in two weeks.

Meta made a statement it hasn’t made in a long time. After launching on 8th September, the company’s new personal AI agent Muse has already achieved more than 900,000 downloads in its first week, which made it the No.1 App on the iOS App Store and Google Play in the US. The number of downloads grew to between 2.5 and 2.8 million in two weeks.

 

Wall Street noticed. Its stock surged by 11% after it was launched, marking the highest one-day increase since April 2025.

 

What Muse actually does

Muse is different to the other chatbot assistants that people are familiar with; it’s proactive. It doesn’t simply provide solutions to queries. Performs multi-step tasks independently: sending email, making travel arrangements, managing personal objectives and interacting with third-party applications.

The agent is installed in a virtual machine environment called “Muse Secure VM,” which is dedicated to the secure storage of user data. It integrates with Meta’s internal services, such as WhatsApp, Instagram and Facebook, as well as external ones such as Gmail and Spotify.

Sentinel’s supervisor system and one-time payment cards for transactions are among the security features.

 

The business model and early benchmarks

Meta organised Muse and will offer two paid plans: $20 per month in the Power plan and $100 per month in the Maximum plan.

During its early rollout, early performance metrics indicate that Muse has also shown to be better suited to certain deployment metrics when compared to ChatGPT, although the specific comparisons have not been clearly outlined.

The launch is being seen as proof of the $200 billion that Meta has reportedly dedicated to AI.

 

Meta’s AI infrastructure play

Muse is no flash in the pan. Over the past several years, Meta has been reorganising its AI business to create a dedicated research arm dubbed Superintelligence Labs, which will be tasked with advancing frontier AI capabilities. The Muse agent is built using the Muse model family, Muse Spark, which is a set of AI models specially designed for agentic tasks and not simply licensed from chatbot architectures.

While the core of Muse is focused on helping users achieve personal goals in various aspects of their lives, CEO Mark Zuckerberg describes it as the logical next step in Meta’s mission to connect people.

 

The trust question

One-time payment cards keep the exposure of financial data limited in case of a breach and provide users with a real-time sense of control over their finances.