Claude Duet Explained: The AI Agent That Messages You First (and How to Get In)

Anthropic’s Claude Duet is a proactive agent that briefs you each morning and works in the background. Here is what it does, its guardrails, and how to join the waitlist.

Reviewed by Mindloop Editorial on . Every link in the Sources section at the end was re-checked that day.

Graphic of a chat window showing a proactive AI agent message, illustrating Anthropic's Claude Duet explainer.

Most AI assistants sit there like a polite dog waiting for a command. Claude Duet is built to do the opposite: it messages you first.

Anthropic announced Duet on 21 October 2026 as a research preview, and as of 15 November 2026 it is still handing out access through a waitlist. Here is what it is, what it will and will not do with your money and inbox, and how to get a place in line.

So what is Claude Duet, exactly?

Anthropic calls it “a new app where Claude works as a proactive agent.” An agent is an AI that takes actions for you, not just answers questions. “Proactive” means it brings things to you instead of waiting.

Per Anthropic, Duet “comes to you with what needs your attention, completes tasks you ask of it, keeps working even when you step away, and checks with you before it acts on your behalf.” The company says it built Duet for professionals and small business owners with more on their plates than they can track.

It was designed around Claude Opus 5.5, Anthropic’s model that, according to the company, leads in agentic coding and knowledge work. Anthropic has since released newer models (see our Claude Fable 5.2 explainer), but Duet’s own pages still name Opus 5.5.

What Duet does all day

Anthropic’s launch page describes a few core jobs:

  • A morning brief. Each morning it sends a short summary of what matters across your connected apps, then stays quiet until something needs you.
  • Monitoring. It watches your accounts during the day and messages you when something needs attention.
  • Bigger tasks. Anthropic’s examples include scheduling meetings, building dashboards, creating financial or sales forecasts, prototyping and building web apps.
  • Background work. Duet has “a virtual machine with its own browser,” so your laptop or phone does not need to stay on. You get a notification when a task finishes or needs review.

It plugs into the Claude Marketplace, which Anthropic says holds more than 6,000 connectors and plugins. Named examples include Gmail, Slack, QuickBooks, Stripe, Gusto, HubSpot, DocuSign and Shopify.

Your memory, skills and connected apps from regular Claude carry over, and you talk to it through one ongoing thread, by typing or voice.

The part where it asks permission

An assistant that acts alone sounds great until it emails your landlord. Anthropic’s design leans on approvals.

Duet reads, drafts and prepares work on its own. But 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,” the company says.

Some things stay off-limits even if you say yes. Anthropic’s examples: “trading securities or moving money between accounts.”

Spending works through Stripe’s Link wallet. Duet confirms the item with you, Link asks you to approve the spend, and it pays with a virtual card so your real card number is not shared. It can also check out with Shop Pay. If agents that pay interest you, our ChatGPT Wallet explainer covers OpenAI’s version of the idea.

Three details worth knowing

  1. It checks in every few hours, not constantly. Anthropic says Duet “checks your connected accounts every few hours, because checking nonstop would use up your plan’s limits much faster.” Background work counts against your usage limits like everything else.
  2. It claims not to see your passwords. Anthropic says its credential handling follows the design of Claude Managed Agents, “so it never directly sees your passwords or sign-in tokens.”
  3. It has a built-in tripwire for secrets. A dedicated classifier checks each message and is designed to reject requests to share the strictest tier of information, such as credit card numbers and government ID. Health and personal finance details are shared only when you explicitly tell Duet to send them.

Anthropic says red teams and an external firm penetration-tested the product, and that it falls under the company’s public bug bounty program.

What it is bad at (Anthropic says so itself)

Anthropic is candid here. It says Duet “might misread a thread or miscalibrate what’s urgent,” and that because the core functionality is built around work, “personal errands, such as booking tickets, aren’t quite as strong.” It also cannot yet show you its browser as it works.

TechCrunch, which covered the launch on 22 October, also flagged the errands and browser limits, and listed Meta Muse and OpenAI Dots among the rivals in this new category of proactive agents.

How to try it: the waitlist, step by step

You cannot download Duet and start using it today. Access runs through a queue.

  1. Check your plan. Anthropic’s pages say the preview is open to US users on Pro, Max and Ultra. Some press coverage, including 9to5Mac, mentions only Pro and Max. Team and Enterprise plans are not included yet.
  2. Join the waitlist. Go to claude.com/product/claude-duet and click “Join the waitlist.”
  3. Wait for the email. Anthropic says it will email you “as soon as Claude Duet’s ready for you.” 9to5Mac reports invitations go out in waves. Neither source gives a date.
  4. Open the app. Duet is a standalone app on iOS and Android, and it is also available at claude.ai on the web.

There is no separate price. During the preview, Duet is included in your plan and shares the same usage limits as the rest of Claude. Anthropic notes that larger jobs, such as building a financial model or a slide deck, take more model turns and background work.

What to watch next

Anthropic says it is still testing Duet’s judgment about when to message you, and that you can tell it what you want to hear about and how often. It also lists sign-in through 1Password saved logins as “coming soon.”

Related: if you want Claude in your calendar rather than in your pocket, see how Claude Meetings works in Google Meet calls.

The real test is a boring one: whether a morning brief saves you time or becomes one more notification. Anthropic itself calls this an early version, so expect rough edges.

Sources

Links last checked on 15 November 2026.

Written by Mindloop Editorial

The Mindloop Experience editorial team writes clear, carefully sourced explainers and how-tos on AI and technology, software and the business of tech, for anyone who likes to know how things actually work.

Meet the editorial team

Responses

Write a response

More from Mindloop Experience

See all from Mindloop Experience

You can clap up to 50 times.