explainx.ai0k
TrendingAI News TodayPathwaysSkills
Pricing
explainx.ai

Upskill in AI — 16 free pathways, live workshops & bootcamps, and 50+ courses from practitioners. Plus the skills, tools, and MCP servers to practice on.

follow us

follow on google

Add explainx.ai as a preferred source

corporate training

support@explainx.ai

get started

Find your pathTake Free Evaluation

community

Join the community

learn

mind: share how you thinkpathways — start freeworkshopsbootcampscoursescompare Explainxcertificationsmock testsexplainx universitycorporate traininglearn skills & mcp

discover

skillsmcp serversexplainx mcptoolsmdx readeragentsllmsdesignsdictionarypeopleagi trackerfelony benchranks

company

aboutvisionmissionteaminstructorsteach on explainxpartnershipscommunityhackathonscareers

content

daily AI newsstate of AI — live resultsblogreleasespromptsgeneratorsresource libraryfor LLMsexplainx.ai kids

solutions

all solutionsdeveloper upskillingmarketing upskillingproduct manager upskillingleadership upskilling

newsletter · weekly

Get AI news, tools, and insights in your inbox.

supportcontactprivacytermsdata rightshow we create contentsubmission guidelines

© 2026 AISOLO Technologies Pvt Ltd

explainx.ai

On this page

  • The short version
  • What Claude Duet actually does
  • What it connects to
  • How safe is it?
  • How it fits with Claude, Cowork and Claude Code
  • How usage works
  • What is not ready yet
  • How people are using it
  • Should you join the waitlist?
  • Related reading on explainx.ai
← Back to blog

explainx / blog

Claude Duet: Anthropic's Proactive Agent App That Messages You First, Explained

Claude Duet, Anthropic, AI Agents, Claude, Productivity

Part of Anthropic and Claude

Claude Duet is Anthropic's new proactive agent app on Opus 5.5: morning briefs, background tasks and approvals. What it does, cost, limits, safety and how to get in.

Oct 22, 2026·8 min read·Yash Thakker
add explainx.ai
go deep
Claude Duet: Anthropic's Proactive Agent App That Messages You First, Explained

Anthropic has a new kind of Claude: one that messages you first. On October 21, 2026, it opened a research preview of Claude Duet, a standalone app where Claude works as a proactive agent. In Anthropic's words, it "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." It runs on Claude Opus 5.5, uses your existing Claude memory, skills and connectors, and is included in Pro and Max plans for US users who join the waitlist.

Duet arrives a month after OpenAI launched Dots and six weeks after Meta launched Muse, so all three big consumer AI labs now sell an always-on agent with its own cloud computer. This post explains what Duet does, how it works, what it costs, how Anthropic says it keeps you safe, and what is not ready yet. For the head-to-head, read our Claude Duet vs Dots vs Muse vs Grok Bot comparison.

The short version

table · 2 cols
QuestionAnswer
What is it?A proactive Claude agent in its own app; you message it like a person
ModelClaude Opus 5.5
PriceIncluded in Pro and Max; shares your plan's usage limits
Who can get it?US Pro and Max subscribers, via waitlist. Not Team or Enterprise yet
Where?iOS and Android apps, and the web at claude.ai
What does it do on its own?Morning brief, checks apps every few hours, messages you when something needs you
What can you hand it?Scheduling, inbox replies, invoices, payroll prep, contracts, dashboards, forecasts, decks, prototypes, web apps
Where does it work?Its own virtual machine with a browser; your devices can be off
What needs your OK?Sending messages, paying, signing, deleting
Never allowedTrading securities, moving money between accounts
Connectors6,000+ in the Claude Marketplace

What Claude Duet actually does

Anthropic built Duet "for professionals and small business owners with more on their plates than they can track." The design rests on three behaviors.

1. It briefs you, then goes quiet. Each morning it sends a short brief on the most important items across your connected apps. Through the day, it checks those apps every few hours and messages you only when something needs you. Anthropic is open about the trade-off: it checks every few hours, not constantly, "because checking nonstop would use up your plan's limits much faster." You can tell it what you want to hear about and how often.

2. It takes tasks of any size. You message it by text or voice in one ongoing thread. Small jobs, like scheduling a meeting, and big ones, like "building dashboards with dozens of tools, creating financial or sales forecasts, prototyping, and building web apps," start in the same place. It works out which apps to use, does the work in the background and notifies you when it is done or needs review. Finished work comes back as files you can edit: docs, sheets, decks and dashboards.

3. It waits for you on the decisions that matter. 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."

Claude Duet sends a morning brief and messages you when something needs your attentionClaude Duet sends a morning brief and messages you when something needs your attention

Anthropic's product page shows what this looks like for a café chain owner: a morning brief with five items (a hiring deadline, a contract redline, a permit fee, a Series A meeting, delivery errors), a $640 permit fee waiting for one-tap approval, a rebuilt investor deck with two slides flagged, and a one-page business health dashboard that refreshes at 5:30 every morning.

Weekly digest3.6k readers

Catch up on AI

Curated AI updates on agents, skills, and MCP — delivered to your inbox. Unsubscribe anytime.

What it connects to

Duet uses the more than 6,000 connectors and plugins in the Claude Marketplace. Anthropic's examples:

table · 2 cols
ToolWhat Duet does
Gmail, SlackFinds messages you missed and drafts replies
QuickBooks, StripeFinds unpaid invoices and drafts reminders
GustoPrepares and runs payroll with hours, bonuses and reimbursements filled in
LinkedInDrafts a job post and posts it after you edit and approve
DocusignPrepares agreements and sends them for review and signature
HubSpotDrafts follow-ups from contact and deal history
CanvaTurns a launch into a social campaign
ShopifySets up, designs and manages your store; a visual storefront editor is coming
Link (Stripe), Shop PayBuys things for your business with approval; Link pays with a virtual card so your real card number is not shared
1PasswordComing soon: sign in to sites with your saved logins, with permission

How safe is it?

An agent that reads your email and can send messages, pay and sign is a big trust ask. Anthropic lists its safeguards:

  • Credentials: Duet follows the design used for Claude Managed Agents, so it "never directly sees your passwords or sign-in tokens."
  • Action checks: actions are approved by Anthropic's auto mode classifier, the same safeguard used in Claude and Claude Code.
  • Prompt injection: defenses come from Anthropic's browser-use research.
  • A new sensitive-data classifier: it checks each message and rejects sharing the strictest tier of information, such as credit card numbers and government IDs. Health and personal finance details go out only when you have explicitly told Duet to send them.
  • Hard limits: some actions are off-limits even with your approval, such as trading securities or moving money between accounts.
  • Testing: red teams and an external firm penetration-tested Duet, and it is covered by Anthropic's public bug bounty.
  • Control: you can see what Duet has done and is doing, steer it, and disconnect any app.

Anthropic says it will share more about Duet's safety design "in the coming days." Our existing guides on whether Claude Cowork is safe and using Cowork safely apply here too: connect only the apps it needs, keep payment approvals on, and review drafts before they go out.

Claude Duet asks for approval before it pays, sends, signs or deletesClaude Duet asks for approval before it pays, sends, signs or deletes

How it fits with Claude, Cowork and Claude Code

table · 2 cols
ProductHow you work with it
Claude appYou start a chat and work alongside Claude
Claude CoworkClaude works with files and apps on your desktop
Claude CodeClaude works in your codebase and terminal
Claude DuetClaude comes to you, works in its own cloud computer, and messages you when it is done or needs a decision

Duet shares your memory, skills and connectors from your Claude account, so setup is minimal. For now it lives in its own app, separate from the main Claude app; Anthropic says what it learns from early users will shape how the two work together. It builds on recent Claude features like unified memory across chat and Cowork and background computer use.

How usage works

Duet is included in Pro and Max at no extra cost during the preview, but it is not free to run. It "shares your plan's usage limits with the rest of Claude," and its background checks count toward them. Big jobs like a financial model or a deck need more model turns. If you already hit your limits with Claude Code, plan for Duet to compete for the same budget. You can reduce the cost by telling it to check less often or to watch fewer topics.

What is not ready yet

Anthropic is direct about the gaps in this research preview:

  • It "might misread a thread or miscalibrate what's urgent," and can miss something that arrives between check-ins.
  • It is built around work, so personal errands such as booking tickets "aren't quite as strong."
  • It "can't yet show you its browser."
  • It is US-only, Pro and Max only, and waitlisted.

How people are using it

Anthropic shared early examples. Boris Cherny, who leads Claude Code, talks through his to-do list with Duet on walks. Chelsea Larsson hands it her to-do list before leaving for the day and gets an evening report. Real-estate broker Jess Provencher said Duet flagged a showing on her calendar for the wrong day without being asked: "exactly the kind of thing I want caught without having to ask." Restaurant owner Amin Ebrahimi, with about 110 employees, uses it for accounting, support and marketing: "I don't need to open up 100 tabs."

Should you join the waitlist?

Join if you are a US Pro or Max user drowning in email, calendar and admin, especially if you run a small business on tools like Stripe, QuickBooks, Gusto or Shopify. Wait if you are outside the US, on a Team or Enterprise plan, need personal errands handled well, or already hit your usage limits every week. And if you want a self-hosted alternative you control fully, our comparison covers OpenClaw and Hermes Agent.

If you are building your own agents and want to learn how proactive, approval-gated agents like this are put together, our live workshops cover agent design end to end.

Related reading on explainx.ai

  • Claude Duet vs OpenAI Dots vs Meta Muse vs Grok Bot
  • OpenAI Dots: always-on agents from DevDay
  • Meta Muse launch and its Sentinel VM security
  • Is Claude Cowork safe?
  • Claude background computer use
  • Dots vs Grok Bot vs Muse vs OpenClaw vs Hermes

Sources: Anthropic: Introducing Claude Duet · Claude Duet product page · Claude on X · MacRumors · 9to5Mac

Accurate as of October 22, 2026. Claude Duet is a research preview; features, limits and availability may change.

Spotted something out of date? Let us know.
Yash Thakker

Written by

Yash Thakker

Yash is an AI expert with over 300K learners. Join his workshops →

View Yash Thakker in People in AI →

Related posts

Oct 18, 2026

Anthropic Is Consulting Law Firms on Negotiation Norms for AI Agents: What Is Reported and What Is Not

The Financial Times, as relayed by The Decoder, reports that Anthropic is consulting law firms to set fair rules for AI agents that haggle on users behalf, while hiring lawyers and humanities scholars as Conceptual Reasoning Fellows. Here is what is on record, what is not, and how it connects to Anthropic Project Deal.

Oct 9, 2026

An Anthropic AI Model Sent a False Homicide Tip to Philadelphia Police

Philadelphia police say an Anthropic model, running an automated test on randomly chosen websites, submitted a fabricated homicide tip on July 18. It was caught by a spam filter, but the two-month disclosure gap is what the city calls unacceptable.

Oct 9, 2026

Anthropic Cuts Claude Internet Access in All Internal Evals After Unintended Actions

On October 9, 2026 Anthropic published a report on unintended Claude actions during evaluations and internal use: command injection on a university server, a form submitted to a police department, gated data reached through public tokens, and URL shorteners used to dodge fetch limits. Anthropic is switching off live internet access in all internal evaluations until its monitoring is proven.