Anthropic has opened a research preview of Claude Duet, a new app where Claude acts as a proactive agent instead of waiting for your next prompt. Claude Duet watches your connected apps, sends you a morning brief, works on tasks in the background, and asks for your approval before it sends, pays, signs or deletes anything. It starts with Pro and Max subscribers in the US through a waitlist.
What Claude Duet Actually Does
Most chatbots sit idle until you type something. Claude Duet flips that model, because it reaches out to you when something needs a decision. Anthropic says it built the app for professionals and small business owners who juggle more than they can track.
You talk to it like you would text a coworker, by typing or sending a voice message. Your main thread stays in one place, so a quick question and a weeklong project start in the same conversation. You can also open separate channels for different topics, and they share the same memory.
A Daily Brief and Ongoing Check-Ins
Each morning, Claude Duet looks across your email, calendar and other connected apps and picks the items that need you most. It can start the first steps, such as drafting replies for you to review. During the day it checks your accounts every few hours and only messages you when something needs attention.
Anthropic says it chose “every few hours” on purpose, since constant checking would burn through plan limits much faster. You can also tell it what you want to hear about and how often.
Bigger Jobs That Keep Running
Claude Duet runs on its own cloud virtual machine with its own browser. That means your laptop or phone doesn’t need to stay on while it works. Finished work, such as a research brief, a financial model, a slide deck or a dashboard, comes back as files you can review and edit.
Claude Duet Availability and Pricing
The preview is invite-only for now, and invitations go out in waves. Here is a quick look at the launch details Anthropic shared.
| Detail | What Anthropic Says |
|---|---|
| Status | Research preview, waitlist only |
| Who can join | Pro and Max users in the US |
| Team and Enterprise | Not supported yet |
| Platforms | iOS app, Android app and claude.ai on the web |
| Extra cost | None, but it uses your plan’s usage limits |
| Model | Built around Claude Opus 5.5 |
| Connectors | Over 6,000 connectors and plugins in the Claude Marketplace |
One catch is worth knowing up front. Background checks and large tasks count toward your limits, so heavy use could push you to your cap sooner than regular chat would. If you already use the Claude AI assistant a lot, keep an eye on that.
Safety Rules and Approvals
An agent that reads your email and acts for you needs firm limits, and Anthropic leans hard on that point. Claude Duet reads, drafts and prepares work on its own. Before any action with real consequences, it shows you exactly what it plans to do and waits.
- It asks before it sends a message, makes a payment, signs a document or deletes anything.
- Some actions stay off-limits even with your approval, such as trading securities or moving money between accounts.
- Its handling of credentials follows the design Anthropic uses for its managed agents, so it never directly sees your passwords or sign-in tokens.
- A new classifier checks each message and is designed to block sharing of the most sensitive data, such as credit card numbers and government IDs.
- Health and personal finance details go out only when you explicitly tell it to send them.
- Red teams and an outside firm ran penetration tests, and the app is covered by Anthropic’s public bug bounty.
Actions are also screened by the same auto mode classifier used in Claude Code, which is meant to stop steps you didn’t ask for. If you have dealt with Claude Code permissions, the approach will feel familiar.
The real test for a proactive agent isn’t how much it can do alone. It’s how well it knows when to stop and ask you.
Apps It Connects To
Claude Duet carries over your memory, skills and connected apps from your existing Claude account, so setup is light. For sites without a connector, it can use its own cloud browser. Anthropic listed several sample jobs it can handle.
- Find missed Gmail and Slack messages and help you reply.
- Check QuickBooks or Stripe for unpaid invoices and draft reminders.
- Prepare payroll in Gusto with hours, bonuses and reimbursements filled in.
- Draft a job post and publish it to LinkedIn once you approve.
- Prepare agreements and send them for signature through Docusign.
- Write follow-up emails from HubSpot deal history, or turn a launch into a Canva social campaign.
Shopify store setup is supported too, and Anthropic says a richer visual storefront editor is coming. Purchases through Link and Shop Pay are possible, but only with your approval.
What This Means for You
If you run a small business or a busy solo practice, Claude Duet is aimed squarely at you. It could take a lot of inbox triage, invoice chasing and meeting prep off your plate. It also joins a fast-growing field of AI personal assistants that try to act before you ask.
- Join the waitlist early if you are on Pro or Max in the US, since access rolls out in waves.
- Start by connecting only the apps you truly need, and add more once you trust its judgment.
- Watch your usage, because background work eats into the same limits as chat.
- Review every draft before approving it, especially in the early weeks.
Anthropic admits the app is early. It may misread a thread or misjudge what’s urgent, personal errands like booking tickets aren’t as strong yet, and you can’t watch its browser while it works. For now it also lives in a separate app from the main Claude app.
Sources
Primary source: Anthropic’s Claude Duet announcement. Additional reporting from 9to5Mac. For more of our coverage, see the Opus 5.5 launch and the latest AI news updates.
Conclusion
Claude Duet is Anthropic’s clearest move yet from a chatbot you visit to an agent that comes to you. It pairs background work and daily briefs with strict approval rules for anything that sends, pays, signs or deletes. The preview is limited to US Pro and Max users, and it shares their usage limits. How well it judges what truly needs your attention will decide whether it becomes a daily habit. If you get an invite, start small and let it earn your trust.
