Back to Technology briefing Technology

Anthropic's Claude Duet agent links 6,000 apps, asks before paying or sending

Anthropic opened a waitlisted preview of Claude Duet, an always-on AI agent that watches your email, calendar and business apps, works in the background and asks before it sends, pays, signs or deletes anything.

By US Brief desk · Updated 2026-10-21T21:10:00-07:00

AI-assisted · US Brief desk · Sources listed below

Two men stand side by side in a bright office lobby with wooden booth seating and an indoor tree, during UK Science Secretary Peter Kyle's visit to Anthropic in San Francisco

What happened

Anthropic on Wednesday opened a research preview of Claude Duet, a separate app for iPhone, Android and the web where its Claude AI works as a proactive agent. Instead of waiting for a prompt, Duet sends a short morning brief on what matters across a user's connected apps, checks those apps every few hours, and messages the user when something needs attention. Users can also hand it tasks by text or voice, and it keeps working on its own cloud computer after they close the app.

Why it matters

Duet is Anthropic's entry in a fast-moving race to sell AI that does work on its own rather than just answering questions. Reuters noted that OpenAI launched always-on "dots" agents last month and that Meta's Muse agent drew millions of downloads. The Verge framed Duet as a direct competitor to both. For small businesses, the pitch is that one assistant watches the inbox, the books and the calendar at once; the risk is handing an AI access to email, payroll and payments.

More context

Who can get it: Access is by waitlist only, starting with U.S. subscribers on Claude's Pro and Max plans, with invitations going out in waves. Team and Enterprise plans aren't included yet. Anthropic said Duet uses the same usage limits as the rest of a user's Claude plan, and that its background checks and bigger jobs, such as building a financial model or a slide deck, eat into those limits faster than a normal chat.

What it can do: Duet runs on Anthropic's Claude Opus 5.5 model and can use more than 6,000 connectors in the Claude Marketplace. Anthropic's examples include finding missed Gmail and Slack messages and drafting replies, checking QuickBooks or Stripe for unpaid invoices and drafting reminders, preparing payroll in Gusto, sending contracts through Docusign and setting up a Shopify store. It can buy things for a business through Stripe's Link wallet, which pays with a virtual card, and check out with Shopify's Shop Pay.

In Anthropic's words: "Before it sends a message, makes a payment, signs a document, or deletes anything, it shows you exactly what it's about to do and waits for your approval. A few things are off-limits even with your approval, like trading securities or moving money between accounts." The company also said a new classifier is designed to refuse requests to share credit card numbers and government ID, that Duet never directly sees users' passwords, and that red teams and an outside firm penetration-tested it.

How outlets framed it: Reuters led with the agent race, calling Duet a "proactive assistant for professionals and small-business owners" as rivals compete on autonomous AI. The Verge's headline cast it as Anthropic taking on Meta's Muse. Anthropic's own post focused on trust and judgment, saying it built Duet to "decide what needs you, what can wait, and when to check with you first."

What's still unknown: Anthropic called Duet "an early version" that "still might misread a thread or miscalibrate what's urgent," and said it is weaker at personal errands such as booking tickets and can't yet show users its browser. The company hasn't said how many people are on the waitlist, when the preview will widen beyond the U.S. or to business plans, or whether Duet will eventually cost extra. Anthropic said it will share more about Duet's safety design "in the coming days."

6 listed sources

References listed by US Brief; a source count is not a verification score.

Editorial sourcing notes

Features, limits and safeguards described here come from Anthropic's own announcement, product page and help center; they have not been independently tested by US Brief. Anthropic did not announce a separate price. Photos are file photos and do not show the launch.

Like US Brief? You can support it with a tip.