Meta’s Muse AI agent can send emails, book travel, lower bills, fill out forms, and complete purchases for you, taking everyday tasks off your plate so you spend less time on them. Meta introduced Muse on September 8, 2026, as a personal AI agent for people in the U.S., its largest move into consumer AI so far. The agent connects to the apps and services you already use and acts on your behalf, which means it needs access to more of your personal information than a standard chatbot.<\/p>\n
Muse is built to complete tasks rather than only answer questions. According to Meta, it can send emails, book travel, lower bills, fill out forms, create plans, turn recipe reels into grocery lists, send party invitations, and make purchases. Checkout runs through Link by Stripe, which offers purchase protections that can reduce worry about letting an agent buy on your behalf. Integrations with Shopify’s Shop Pay and 1Password are planned to follow.<\/p>\n
The agent keeps working after you leave the app, and Meta says it improves over time by learning from your conversations what matters to you, then making suggestions without being asked. Muse is powered by Meta’s AI model, Muse Spark.<\/p>\n
You choose which apps and services Muse can reach, and you connect them one at a time to keep the opt-in clear. Muse ships with built-in connectors for several services and adds more over time. If a service you want is not supported but offers a public API, Muse can create a connection using credentials you provide. When no API is available, Muse reaches the service through a browser instead. The list of everyday tools it can plug into includes email, calendars, payments, and apps for health and fitness, the smart home, dining, shopping, music, and events.<\/p>\n
Muse is available on the web at muse.ai, through iOS and Android apps, and inside chats on WhatsApp. Meta says it will also come to its AI glasses. The agent is free to use, with subscription plans that begin as your usage grows, which is why it asks for a payment card at signup. Two paid plans are available at launch: Power at $20 per month and Maximum at $100 per month, both offering more capacity for handing off tasks. Meta expects most people to stay on the free tier and says the app includes a usage meter that shows how much of your allowance remains, warns you when free usage runs out, and presents subscription options.<\/p>\n
Meta says Muse runs inside its own dedicated, secure computer with its own browser, called Muse Secure VM, which applies privacy, safety, and security protections to your data. A separate Sentinel agent runs on that same virtual machine but is kept apart from Muse at the system level. The company says this design means Muse cannot see your passwords or payment methods, and that Muse does not share your conversations or data with Meta’s ads systems. Meta published a technical post explaining these claims, which security experts will need to review independently.<\/p>\n
Muse is also built to feel personal. You can give it a name, choose its avatar, and configure how it looks and communicates with you.<\/p>\n
Muse asks for a level of trust that reaches further than social media did, and Meta’s history shapes how that request lands. The company settled with the FTC in 2011 over charges that it made users’ private information public without approval. In 2019 the FTC imposed a then record $5 billion settlement over eight privacy violations, and in 2023 it charged Meta with violating the privacy order that followed. Also in 2019, Meta found a number of user passwords stored in readable formats, and the Cambridge Analytica scandal, in which data belonging to millions of users was collected by a third party without consent, still shapes public opinion.<\/p>\n
The company has faced repeated questions in Congress over how it protects minors. Related lawsuits followed, including an $18 billion multistate settlement reached with 29 states in August, and a New Mexico case over harms to children in which Meta was ordered to pay $942 million. Thousands of personal-injury and school-district cases remain pending against multiple social media companies.<\/p>\n
Muse arrives as many companies test how AI agents fit into daily life. Some build AI web browsers or task services such as Gemini Spark from Google or Claude Cowork from Anthropic, while others place agents inside the chat apps people use most, including Apple’s iMessage, SMS, and WhatsApp. These agents raise a shared question about how much access users are willing to grant. Early testers of one AI assistant were surprised to find it required a broad, perpetual, and irrevocable license to access, store, reproduce, and modify their materials, including for training AI models. Whether Muse’s utility and its stated protections earn enough confidence for people to connect their accounts is the open question for its adoption.<\/p>\n
Muse can send emails, book travel, lower bills, fill out forms, create plans, turn recipe reels into grocery lists, send party invitations, and make purchases. It connects to apps you already use, such as email, calendars, payments, health and fitness, smart home, dining, shopping, music, and events, and keeps working after you leave the app.<\/p>\n
Muse is free to use, with two paid plans at launch: Power at $20 per month and Maximum at $100 per month, both offering more capacity for handing off tasks. A payment card is required at signup because plans begin as usage grows, and a usage meter shows how much free allowance remains. Meta expects most people to stay on the free tier.<\/p>\n
Meta says Muse does not share your conversations or data with its ads systems. The agent runs inside a dedicated secure computer called Muse Secure VM with its own browser, and Meta says this design keeps Muse from seeing your passwords or payment methods. The company published a technical post on these protections that security experts will need to review independently.<\/p>\n