Facebook’s Muse AI: How the new AI agent works, its features and what it can do

Facebook’s parent company Meta has launched Muse, a personal AI agent designed to go beyond answering questions and instead carry out tasks on a user’s behalf. Meta says Muse can plan, take action and help users manage everyday tasks, marking a shift from traditional chatbots that mainly provide answers.

Introduced on September 8, 2026, Muse is part of Meta’s broader push to develop AI agents that can turn users’ goals into completed work.

Meta describes Muse as an AI agent that can take on tasks and continue working toward a goal.

“Muse is a personal AI agent. It doesn’t just answer questions; it actually does the work,” Meta said.

For now, Muse is rolling out in the United States on iOS, Android and the web through muse.ai, with the company also integrating the experience into WhatsApp. Meta says Muse is coming soon to AI glasses.

What is Facebook’s Muse AI?

Muse comes from Meta, the parent company of Facebook, Instagram, WhatsApp and Messenger. Unlike a conventional chatbot, it is designed to perform tasks and continue working toward a user’s goal.

Meta says the system was built as a personal AI agent that can proactively help people with their goals.

“Today, Meta is introducing Muse, a secure, private personal AI agent that proactively helps with people’s goals and suggests ideas,” the company said.

The company says Muse is designed around the way people already communicate, allowing users to interact with it in a messaging-style format through the Muse app or WhatsApp.

How does Muse work?

Muse is powered by Muse Spark, which Meta describes as its most capable model to date for real-world agentic work.

Rather than requiring users to learn complicated commands, Meta says they can simply describe what they want done.

“People just tell Muse what needs to get done, and it takes action,” Meta said.

Once a user gives Muse a goal, the system can help develop a personalised plan, coordinate time and resources and continue advancing the work.

Meta says Muse can also open a browser, fill out forms and negotiate on a user’s behalf. For longer tasks, it can continue working after the user closes the app and return when something changes or when it requires approval.

What can Muse do?

Muse is designed to handle a range of practical tasks. Meta says it can send emails, book travel, research information, fill out online forms and work across connected services.

The company also says Muse can take on broader goals rather than just individual requests.

“It can handle tasks, like sending an email or booking travel, and it can take on big audacious goals,” Meta said.

Muse can also remember information from previous conversations and use it when making suggestions.

For example, Meta says the agent can take a recipe a user saved on Instagram, turn it into a grocery list, suggest a menu and remember guests’ dietary restrictions when helping organise a dinner.

Muse can work in the background

A major feature of Muse is its ability to continue working after a user leaves the app.

Meta says that for tasks requiring more time, Muse can continue working and return when something changes or when it needs approval. That could include asking the user before sending an email or making a purchase.

This is what separates an AI agent such as Muse from a conventional chatbot. Instead of simply providing information, the agent is designed to take multiple steps toward completing a task.

Shopping with Muse

Muse can also assist with online shopping.

Meta says the agent can check out using Link by Stripe, with Link’s agent wallet generating a one-time-use card so a user’s actual card details remain hidden.

Meta said Muse is the first AI agent covered by Link’s purchase protections, which include protections for eligible purchases.

The company says Shop Pay and 1Password support are also planned.

How Meta says Muse protects user data

Because Muse can interact with online services and potentially access sensitive information, Meta has built security features around the system.

Muse runs on what Meta calls Muse Secure VM, a dedicated virtual machine that houses the agent and a user’s data.

Meta says a separate security system called Sentinel monitors Muse’s actions.

“Nothing Muse does reaches the internet unless the Sentinel approves it,” the company said.

Meta also says Muse cannot see users’ passwords or payment details. Instead, credentials are stored securely so the agent can use them without directly seeing them.

Users can also choose which applications Muse connects to and how much access it receives. Meta says they can change permissions or disconnect a service at any time.