
Claude Duet is a new proactive AI agent app available as a waitlist-based research preview for US Pro and Max subscribers on iOS, Android, and claude.ai. Powered by Claude Opus 5.5, it runs on its own virtual machine with a browser, allowing it to prepare work even when your devices are off. It shows you its planned actions and waits for approval before sending messages, making payments, signing documents, or deleting anything. Anthropic notes that it can misjudge urgency and is less effective at handling personal errands.
Anthropic wants Claude to stop waiting for you to type. On Wednesday, October 21, it opened a waitlist for Claude Duet, a standalone agent app that sends a morning brief, monitors your connected apps, and does tasks in the background. It joins Meta's Muse and OpenAI's dots in a crowded race. Every claim below comes from Anthropic, and no independent testing has been published yet.
Brief you, watch your apps, and take on tasks. According to Anthropic's announcement:
Morning brief: A short summary of the most important items across your connected apps
Background monitoring: It checks connected accounts every few hours and messages you only when something needs attention
Admin tasks: Scheduling meetings, finding missed Gmail and Slack messages, drafting invoice reminders from QuickBooks or Stripe
Bigger jobs: Building dashboards, financial or sales forecasts, prototypes and web apps
Business operations: Preparing payroll in Gusto, sending agreements through DocuSign, drafting HubSpot follow-ups
Purchases: Buying items through Stripe's Link wallet with a virtual card, which you approve, or through Shopify's Shop Pay
You can message it by typing or voice in one ongoing thread. Its memory, skills, and connected apps carry over from your Claude account.
Only a limited group for now. The research preview is open to Pro and Max subscribers in the US, and Anthropic's announcement does not mention Team or Enterprise plans. Pro costs $20 a month, and Max starts at $100. Duet adds no separate price in the preview, but its background checks count toward your plan's usage limits, and larger tasks such as building a deck or a financial model use more of them. Anthropic gave no task-level cost estimates, productivity figures, or benchmark results, according to VentureBeat.
We helped organizations choose the right AI platform, integrate, and scale AI solutions that drive real business impact and track measurable results.
Book a Free ConsultationWith gates for consequential actions and several layered safeguards. Anthropic says:
Approval first: Before it sends a message, pays, signs, or deletes, Duet shows you the action and waits. Trading securities and moving money between accounts are off-limits even with approval
No direct credential access: Duet never sees your passwords or sign-in tokens, using the design behind Claude Managed Agents
A new sensitive-information classifier blocks requests to share the strictest tier of data, such as credit card numbers and government ID. Health and finance details go out only if you explicitly say so
Testing: Red teams and an external firm penetration tested it, and it is covered by Anthropic's bug bounty
Anthropic says it will share more safety details in the coming days.
Anthropic is candid that it is an early version. Duet can misread a thread or miscalibrate what is urgent. It is built around work, so personal errands such as booking tickets are not as strong, and it cannot yet show you its browser. It lives in its own app, separate from the main Claude app, and checking only every few hours is a deliberate usage trade-off. Engadget noted that a heavy agent sharing limits with Claude Code could use up rate limits quickly, and that how much you get from it depends on how much you trust Anthropic with your digital life.
Similar goal, different audience and access. CNBC reports Meta has marketed Muse mainly to consumers, while OpenAI's dots straddle consumer and enterprise use. Engadget says dots are currently limited to OpenAI's Pro tier, starting at $100 a month, whereas Duet is included in Pro and Max. Anthropic positions Duet for professionals and small business owners. Muse reached the top of Apple's App Store after launching in September, and we covered its model family. Engadget called Anthropic "late to the party." All three are early products, so direct comparisons will need independent testing.
Treat it as a preview to evaluate, not a system to roll out.
Set a policy now on employees connecting work email, calendars or payment tools to personal agent apps, since the announcement does not cover Team or Enterprise plans
Test approval flows and usage costs on non-sensitive work before relying on it
Note the product pattern: Agents are arriving as standalone mobile apps with their own cloud machines and approval gates
Duet runs on Claude Opus 5.5, whose pricing we covered earlier. Our mobile application development and agentic AI development teams can help you plan agent-powered apps and workflows with sensible approval and access controls.
Claude Duet is a proactive AI agent app from Anthropic, launched as a research preview on October 21, 2026. It sends a daily brief, monitors your connected apps, completes tasks in the background, and asks for approval before consequential actions. It is built around Claude Opus 5.5, runs on iOS, Android, and claude.ai, and is aimed at professionals and small business owners.
US subscribers on Claude Pro and Max plans can join the waitlist. Anthropic will email people when their access is ready. The announcement does not mention Team or Enterprise plans, and Anthropic says it is starting with US users. Because it is a research preview, access is limited, and the product may change as Anthropic tunes it based on early user feedback.
Anthropic announced no separate charge for Duet during the preview. It is included in Pro ($20 per month) and Max (from $100 per month) plans. However, it shares your plan's usage limits with the rest of Claude, and its background checks and larger tasks count toward them, so heavy use could consume your allowance faster than ordinary chatting.
Yes. Anthropic says Duet has its own virtual machine with a browser, so it does not need your laptop or phone to be on. You describe a task in a short message, and it works out which apps and steps to use, then notifies you when it is finished or needs review. You cannot yet watch its browser.
Not without your approval. Before it sends a message, makes a payment, signs a document, or deletes anything, Duet shows you exactly what it will do and waits for your go-ahead. Some actions, such as trading securities or moving money between accounts, are off-limits even with approval. Purchases use Stripe's Link, which asks you to approve the spend.
No, according to Anthropic. Duet never directly sees passwords or sign-in tokens, following the credential design used for Claude Managed Agents. Anthropic also added a classifier that rejects requests to share the strictest information, such as credit card numbers and government ID, and lets health or personal finance details through only on your explicit instruction.
Mainly in audience and access. Meta markets Muse mostly to consumers, and OpenAI's dots serve consumers and businesses and are limited to its Pro tier, which starts at $100 per month. Anthropic positions Duet for professionals and small business owners and includes it in Pro and Max. All three are new, so independent comparisons of quality and reliability are still needed.
Decide what data and accounts an agent may access, and who approves it. Anthropic's announcement does not mention Team or Enterprise plans, so staff might connect work accounts to personal plans unless you set a policy. Test approvals and usage costs on low-risk work first, and review Anthropic's promised safety details when they are published.
We are more than just developers and consultants—we are your partners in navigating the digital landscape. Let us be the engine behind your next big success while you focus on your core vision.
Explore Opportunities!