MarkGrid

What Is Claude Duet? Anthropic's Proactive AI Agent App Explained: Features, Access, Privacy and Rivals

Anthropic's Claude Duet is a proactive agent you message: morning briefs, inbox and calendar help, approval before sensitive actions. Here's who can use it and the concerns.

KT
Kunal tomarGrowth and content lead
Oct 22, 2026 7 min read
Cover reading "Claude Duet: The agent pings first" on a dark grid with a violet network graphic

Claude Duet is a new Anthropic app, announced on October 21, 2026, where Claude works as a proactive personal agent you message. It watches your connected apps, sends a morning brief, handles inbox, calendar and follow-up work, and runs bigger jobs like decks and dashboards in the background. It asks before sensitive actions and is waitlist-only for US Pro and Max subscribers.

Key facts: Announced October 21, 2026 as a research preview; built on Claude Opus 5.5 and runs on its own cloud virtual machine; works with 6,000+ connectors and plugins; asks approval before sending messages, paying, signing or deleting; waitlist for US Pro and Max, not Team or Enterprise, and uses your existing plan limits.

What is Claude Duet?

Claude Duet is a standalone app in which Claude acts as an agent that reaches out to you, rather than waiting for you to open a chat. The Claude Duet product page describes it as a personal agent you message that handles your inbox, calendar and follow-ups, and contacts you when something needs your input.

Anthropic's launch announcement says it is aimed at professionals and small business owners with more to track than they can manage. Everything happens in one thread, by text or voice.

What can Claude Duet actually do?

Claude Duet works on a daily rhythm: a brief in the morning, check-ins during the day, and longer tasks on request. According to Anthropic, it:

  • Sends a morning brief summarizing the most important items across your connected apps, with suggested next steps.
  • Checks your apps every few hours and messages you only when something needs attention.
  • Takes on larger tasks, such as scheduling meetings, building dashboards, preparing financial or sales forecasts, drafting briefs and decks, prototyping and building web apps, then notifies you when the work is done.
  • Carries over your memory, skills and connected apps from your existing Claude account, so setup is light.

It runs on its own virtual machine with a browser, so work continues while your devices are off. Anthropic's examples include finding missed Gmail and Slack messages, chasing unpaid invoices in QuickBooks or Stripe, and drafting follow-ups from HubSpot deal history. It can also buy things through Stripe's Link wallet and Shopify's Shop Pay. One restaurant owner quoted in the announcement summed up the appeal: "I don't need to open up 100 tabs."

When did Anthropic launch Claude Duet, and who can use it?

Anthropic announced Claude Duet on Wednesday, October 21, 2026, and the launch date is confirmed by the company's own post and by Reuters, Engadget and MacRumors. Access works like this:

  1. Join the waitlist on the product page. It is a research preview, not a general release.
  2. Be a US subscriber on Claude Pro or Max. Team and Enterprise plans are not included yet.
  3. Use the app on iOS, Android or the web at claude.ai, according to Anthropic.

There is no extra fee, but Duet shares your plan's usage limits with the rest of Claude. Background checks count, and larger jobs like decks use limits faster.

Reports differ on details. Engadget described mobile apps as planned and said Pro users get access at general availability, while Anthropic lists iOS and Android apps and both Pro and Max on the waitlist. MacRumors mentions only iOS and the web.

How does Claude Duet handle safety and privacy?

Claude Duet shows you sensitive actions and waits for approval before sending a message, making a payment, signing a document or deleting anything. Anthropic also lists these safeguards:

  • The agent never directly sees passwords or sign-in tokens.
  • A new classifier blocks sharing of the strictest data, such as card numbers and government ID. Health and personal finance details are shared only when you explicitly say so.
  • Trading securities and moving money between accounts are off-limits, even with your approval.

Anthropic is open about the limits. The product page warns it "may misread a thread or miscalibrate what's urgent." Personal errands like booking tickets are weaker than work tasks, and you cannot yet watch its browser.

What are the concerns about proactive agents like Duet?

The main concern is trust: an agent that reads your email, calendar, payments and CRM around the clock holds a detailed picture of your working life. Engadget argued that Duet's value depends on how much users trust Anthropic with that data, and noted that safety measures were described only at a high level so far.

Other criticisms are practical. Background checks eat into the same usage pool as Claude Code and regular chats. Some readers on MacRumors pushed back on the whole idea of an AI that pings you first, calling the push to delegate everything "exhausting." A further risk: frequent approval prompts can train people to approve without reading.

How does Claude Duet compare with OpenAI Dots and Meta Muse?

Claude Duet arrives after two similar products. Coverage from Reuters, Engadget and Digital Trends frames it as Anthropic's answer to both.

  • Meta Muse, announced September 8, 2026, is a personal agent running on a cloud virtual machine with a visible browser and an approval system, available in the US with a free tier, according to Axios.
  • OpenAI Dots, launched September 29, 2026, are always-on agents in ChatGPT with their own cloud computers and access to 4,000+ apps, initially for Pro and Business Premium subscribers, per SiliconANGLE.
  • Claude Duet stands out for its larger connector catalog (6,000+), its focus on work tasks for small businesses, and its inclusion in the $20 Pro tier rather than a $100+ plan.

What does Claude Duet mean for marketers?

For marketers, Duet shows how agents are moving out of separate tools and into the daily workflow itself. Many of the jobs it targets, like a morning roundup of campaign messages, a HubSpot follow-up, a client brief before a call or a weekly dashboard, are routine marketing operations. If you are new to the idea, our explainer on what AI agents in marketing are covers the basics.

A few practical points:

  • Start with read-only work. Briefs, summaries and dashboards carry little risk. Keep anything customer-facing behind the approval step and actually read it.
  • Map which systems you connect. Every connector widens what the agent can see, so check agency and client contracts before linking client accounts.

General-purpose assistants are good at admin, but they do not tell you how your brand appears inside ChatGPT, Claude or Gemini answers, or whether a creative will hold attention. That is where specialist systems such as MarkGrid's Model Share for AI share of voice fit, and our guide to AI marketing agents explains how the two layers work together.

FAQ

Is Claude Duet free?

There is no separate charge. Claude Duet is included in Claude Pro and Max plans during the research preview, but it draws from the same usage limits as the rest of Claude.

Can Team or Enterprise customers use Claude Duet?

Not yet. The preview starts with US Pro and Max subscribers, and Team and Enterprise plans are not included.

Will Claude Duet send emails or pay bills on its own?

No. It shows you the action and waits for approval before sending messages, making payments, signing documents or deleting anything. Trading securities and moving money between accounts are blocked entirely.

What model powers Claude Duet?

Claude Opus 5.5, running on its own cloud virtual machine.

Is Claude Duet available outside the US?

Not at launch. Anthropic has limited the research preview to US subscribers and has not given a date for other countries.

How is Claude Duet different from regular Claude chat?

Regular Claude waits for your prompt. Duet checks your apps every few hours and messages you first.

Last updated October 22, 2026.

Model Share for this topic
ChatGPT
34%
Gemini
28%
Perplexity
41%

How often MarkGrid is named when AI models discuss this topic. About Model Share

KT

Kunal tomar

Growth and content lead

Kunal Tomar is a Growth and content lead at Markgrid

See it on your brand

Thirty minutes on where marketing is leaking for you.

Talk to Sales