Technology

Meta's Muse Tops U.S. App Charts With Fast Early Downloads

Martin HollowayPublished 6m ago4 min readBased on 9 sources
Reading level
Meta's Muse Tops U.S. App Charts With Fast Early Downloads
source:fb.com

Meta's Muse held the No. 1 free-app spot on the U.S. App Store from September 18 through September 25, 2026, and on Google Play since September 19. The highest third-party estimate puts total downloads at 4.3 million. Availability is still limited to the U.S. and Canada.

Meta launched Muse on September 8, 2026, as a free personal AI agent for Mac and mobile. The app runs on Meta's Muse Spark AI model, according to Fortune. Meta describes it as a secure, private personal AI agent that proactively helps people meet their goals and suggests ideas.

Muse is built to act, not only chat. A chat-only tool mainly answers questions. An agent can also complete steps toward a goal. Meta states Muse can answer questions, complete tasks, browse the web, make purchases, generate images, create documents, and connect with apps and services. Meta also states it works across phone, computer and the web.

Muse averaged 55% day-over-day download growth from September 8 to September 17, according to TechCrunch. It had passed 2.5 million downloads by the start of the week of September 25. In the U.S. and Canada, it recorded more downloads and more daily active users than ChatGPT did over a comparable early-launch period.

Third-party totals differ. Apptopia estimated 4.3 million downloads to date, with 2.6 million on iOS alone. Sensor Tower estimated more than 3.4 million downloads as of its Thursday read, including 2.8 million in the first two weeks on app stores. Appfigures estimated roughly 2.3 million installs as of Thursday, split almost evenly between iOS and Android. The spread reflects different data panels, estimation models and methods for removing duplicates.

Meta is marketing the Muse app to existing Facebook and Instagram users. Sensor Tower data indicates Meta started house ads promoting Muse on September 9, a day after launch. House ads are ads shown in a company's own apps. That owned inventory lowers acquisition cost and lets a new app start from billions of existing logged-in sessions.

Sensor Tower found daily active users for the Muse app climbed 27% on Wednesday after Meta Connect. At Connect, Meta announced upcoming Muse features including video chat with the Muse avatar, computer use on the Mac, a dedicated email address, more partners and connectors, and smart glasses integrations. Connectors are links that let the agent work inside outside apps and services. Those items are still roadmap, not shipped capability.

Meta describes Muse Charm as its first AI agent device, and positions Muse as spanning glasses alongside phone, computer and web.

The broader context here is a shift in what the app is for. Chat established the prompt box. Muse is framed around delegation, memory of goals over time, and direct action inside third-party services through connectors, plus computer use on macOS and a dedicated email address for agent-led work. That design raises familiar questions around permission scoping, credential storage, purchase authorization, and audit trails for actions taken on a user's behalf.

In my view, the early numbers say less about model preference than about distribution and timing. Topping the App Store ahead of OpenAI's ChatGPT, as reported by Fox Business, is notable. Doing so in only two countries, with in-house promotion and a Connect news cycle behind it, suggests headroom if and when geographic expansion follows. Worth flagging is retention. Installs compounded at 55% daily for nine days, and daily active use rose after Connect, but sustained use of delegated tasks is harder to earn than a first-week trial.

What this enables, if the connector and computer-use pieces land as described, is a test of agents as daily utilities rather than demos. Purchases, documents, image generation and web browsing inside one agent, reachable by voice, text, avatar video and glasses, would move more routine coordination work off manual app-hopping. That is the practical prize, and the reason the next data to watch is not rank charts but repeat task completion.