Pull to refresh
Logo
Meta launches Muse, an AI agent that acts across personal apps

Meta launches Muse, an AI agent that acts across personal apps

New Capabilities

The assistant runs in its own secure cloud computer and can send emails, book travel, and negotiate for users

4 days ago: Muse launches in the US

Overview

Updated 3 days ago

Meta released Muse, a personal AI agent, to U.S. users 18 and older on September 8. The agent runs its own browser inside a dedicated cloud computer and can send emails, book travel, fill out forms, and negotiate on a person's behalf.

Muse is the centerpiece of Meta's attempt to catch OpenAI and Anthropic in the race to build agents that act rather than just answer. It works through the Muse app or WhatsApp, keeps running after users close the app, and pauses to ask permission before sensitive actions like sending a message or making a purchase.

The launch comes weeks after Meta agreed to a multibillion-dollar settlement with state attorneys general over consumer harms. The company is betting that users will trust an agent with access to their email, calendar, and payment methods.

Why it matters

If Muse earns trust, it becomes the default gateway for shopping, payments, and scheduling, handing Meta control over how users act online.

Questions about this story

Free account needed to ask — your question is kept and asked for you right after sign-up. Answers are public.

No questions yet — be the first to ask.

Key Indicators

18+
Minimum user age at launch
US users must be at least 18 to access Muse.
2
Launch access points
Users reach Muse through its own app and inside WhatsApp.
1
Dedicated secure VM per user
Each agent runs in an isolated Muse Secure VM, separated from other users' agents.

Voices

Curated perspectives — historical figures and your fellow readers.

Ever wondered what historical figures would say about today's headlines?

Sign up to generate historical perspectives on this story.

People Involved

Organizations Involved

Timeline

3 events Latest: 4 days ago
  1. Muse launches in the US

    Latest Product Launch

    Meta releases the personal AI agent to US users 18 and older through the Muse app and WhatsApp.

  2. Reuters reports pre-launch test failures

    Report

    Reuters says internal tests showed Muse stalling and exposing sensitive data without authorization.

  3. Meta details Muse security architecture

    Statement

    Meta describes the Sentinel agent that gates internet access inside each user's secure virtual machine.

Scenarios

1

Muse becomes a daily habit for millions

Possible Resolves by Sep 8, 2027

Discussed by: Bloomberg and The Verge frame Muse as Meta's bid to put agents in mainstream hands

If the agent performs reliably and users accept the secure-VM model, adoption could compound quickly. WhatsApp's reach of billions gives Muse a distribution pipeline rivals lack. Growth would show up in Meta's user metrics and usage disclosures over its first year.

2

A security failure stalls the rollout

Possible Resolves by Q1 2027

Discussed by: Reuters documented stalling and sensitive-data exposure in pre-launch tests; WIRED and TechCrunch question consumer trust

Reuters reported internal tests in which the product stalled and exposed sensitive data. If a similar failure surfaces in the wild, regulators and users could force Meta to pause expansion or restrict what Muse may do without approval. The stakes are high because the agent holds access to email, calendar, and payment methods.

3

Muse stalls without broad adoption

Possible Resolves by Sep 8, 2027

Discussed by: TechCrunch and CNBC note the trust hurdle after Meta's recent legal settlement

Muse faces a credibility problem: the company settled a multibillion-dollar consumer-safety lawsuit with states shortly before launch. Sceptical users may decline to connect email or payment accounts, leaving Muse a niche feature. A stall would show as flat usage and a US-only footprint a year after launch.

Historical Context

3 moments from history that rhyme with this story — and how they unfolded.

1997

Microsoft Clippy (1997)

Microsoft shipped Clippy, an animated assistant, with Office 97 to proactively help users write documents. Users found its unsolicited interruptions more annoying than useful, and Microsoft removed it in later versions.

Then

Clippy became a punchline and a symbol of software that gets in the way.

Now

It taught makers that proactive agents must earn the right to interrupt, a lesson still relevant to agent design.

Why this matters now

Muse is proactive by design, making the reception of unwanted suggestions a live question about user tolerance.

October 2011

Apple Siri (2011)

Apple shipped Siri with the iPhone 4S as the first mainstream consumer AI assistant. Launch ads showed it booking restaurants and managing calendars, but real-world performance was limited to narrow scripted queries.

Then

Siri became a phone staple but never delivered the autonomous organizing its launch promised.

Now

It set the template for assistant hype that outran delivery, a pattern repeated across the industry.

Why this matters now

Muse faces the same gap between demoed autonomy and real reliability, but with far more consequential access to payments and communications.

November 2014

Amazon Alexa (2014)

Amazon launched Alexa with the Echo, betting a voice assistant in the home would create new shopping habits. Alexa reached tens of millions of homes but never became a dependable revenue engine.

Then

Alexa popularized consumer voice assistants and dominated the category for years.

Now

Amazon struggled to monetize it and cut jobs in the division, showing that reach does not guarantee profit.

Why this matters now

Muse's economics are unproven, and the historical record suggests adoption hurdles are as much about revenue as about technology.

Sources

(11)