New to this? Read it in simple words
- Claude Duet is a new Anthropic app that does work tasks for you in the background.
- It sends a short brief each morning and messages you when something needs attention.
- Anthropic says it asks before it sends messages, pays or deletes anything.
- For now, only Pro and Max subscribers in the US can join the waitlist.
- AI agent
- An AI program that can take steps on its own, such as booking a meeting, not only answer questions.
- Virtual machine
- A computer that runs inside a data centre, so it keeps working when your own device is off.
- Research preview
- An early version that a company opens to some users, to learn and improve before a full launch.
Anthropic makes the Claude AI models and apps. On Wednesday, October 21, it opened a research preview of Claude Duet, a new app where Claude works as a proactive agent. It comes to you with what needs attention, instead of waiting for questions.
What Duet does
Claude Duet is a separate app for iOS and Android, and it also works on claude.ai. It runs on Claude Opus 5.5, Anthropic’s top model.
Each morning, Duet sends a short brief on the most important items in a user’s connected apps. During the day it checks those apps every few hours and sends a message only when something needs attention, Anthropic says.
Users can also hand it tasks, from scheduling meetings to building dashboards, sales forecasts and simple web apps. Anthropic says it built Duet for professionals and small business owners.
From Anthropic’s announcement of October 21.
Working while you are away
Duet has its own virtual machine with a browser, so it does not need the user’s laptop or phone to be on. It takes over the memory, skills and connected apps from the user’s Claude account.
It can use more than 6,000 connectors and plugins in the Claude Marketplace, such as Gmail and Slack. Anthropic says Duet checks apps only every few hours, because nonstop checking would use up a plan’s limits much faster.
Duet shares its usage limits with the rest of Claude. Users may reach those limits sooner than with chat alone, 9to5Mac reports.
What Anthropic promises on safety
Duet reads, drafts and prepares work on its own. Before it sends a message, makes a payment, signs a document or deletes anything, it shows the step and waits for approval, Anthropic says.
Some actions stay blocked even with approval, such as trading securities or moving money between accounts. Anthropic also says Duet never directly sees passwords or sign-in tokens.
A new filter is designed to refuse sharing credit card numbers and government ID numbers. Health and money details go out only when the user tells Duet to send them, the company says.
Anthropic admits that Duet may misread a thread or misjudge what is urgent. It is also weaker at personal errands, such as booking tickets.
Who can use it
For now, there is a waitlist for Pro and Max subscribers in the United States. Duet is not yet available on Team or Enterprise plans.
During the preview it costs nothing extra on top of the Pro and Max plans. Engadget notes that OpenAI’s dots are only on plans that start at $100 a month, while Claude Pro costs $20 a month.
Meta’s Muse agent launched in September and OpenAI introduced dots weeks later, CNBC reports. Anthropic says it will share more about Duet’s safety design in the coming days.
Sources
Every fact in this story comes from the sources below. Open them to check our work.
- 1
- 2
- 3
- 4Research · October 21, 2026Anthropic announces preview of Claude Duet, a personal AI agent to take on Muse, Dots CNBC
- 5
- 6
What Duet does, its safety rules and its weak spots come from Anthropic itself. No outside group has tested the app yet.
Anthropic launches a new always-on work agent, but it’s a waitlisted preview whose safety promises no outside group has tested. How we rate the mood