Claude Duet
claude.com
October 21, 2026
8 min read
44/100
Summary
Anthropic has opened a waitlisted research preview of Claude Duet, a proactive AI agent for US Claude Pro and Max subscribers. Available as a standalone iOS and Android app and through claude.ai, Duet carries over a user’s Claude memory, skills and connected apps. It provides morning briefs, monitors connected accounts every few hours, alerts users to items needing attention, and can continue assigned tasks in a virtual machine when the user is away. Built around Claude Opus 5.5, Claude Duet can schedule meetings, prepare dashboards and forecasts, prototype web apps, manage business workflows, and use more than 6,000 Claude Marketplace connectors and plugins. Supported tasks include drafting Gmail and Slack replies, identifying unpaid invoices in QuickBooks or Stripe, preparing Gusto payroll, posting approved LinkedIn job listings, and managing Shopify stores. It can make purchases through Link or Shop Pay after confirmation and payment approval. Duet requires approval before sending messages, making payments, signing documents or deleting data; securities trading and account-to-account transfers are prohibited. Anthropic says Duet never directly accesses passwords or sign-in tokens and uses classifiers intended to block sharing credit-card numbers and government IDs. The company warns that the early version can misread threads or misjudge urgency, is less capable for personal errands, cannot show its browser, and shares usage limits with other Claude features.
Key Takeaways
- Claude Duet is a waitlisted research preview for US Claude Pro and Max subscribers, available on iOS, Android and claude.ai.
- Claude Duet monitors connected accounts, sends morning briefs and can complete background tasks through a virtual machine without the user’s device remaining on.
- Claude Duet requires user approval before sending messages, paying, signing documents or deleting data, and it cannot trade securities or transfer money between accounts.
- The service can use more than 6,000 Claude Marketplace connectors and plugins for workflows involving tools such as Gmail, Slack, QuickBooks, Stripe, Gusto, LinkedIn, Docusign, HubSpot, Canva and Shopify.
- Background monitoring and large tasks consume the same Claude plan usage limits as other Claude features.
What the discussion said
The thread barely discussed Claude Duet’s announced functionality itself. Instead, it zeroed in on the familiar gap between a strong underlying model and the product wrapped around it. One reader expressed confidence in Claude’s model quality but worried that a buggy application could squander that advantage, arguing that a competing assistant has a more polished non-model experience. That supposed competitor advantage immediately drew pushback: another reader found its web search and information-gathering useful, but judged its memory, construction ability, and generated artifacts too weak for serious work. The exchange suggests that product reliability and persistent context matter as much as raw model intelligence when assistants are meant to produce usable outputs. A separate comment shifted to privacy architecture, calling for these increasingly capable assistants to run with trusted-execution-environment attestation. The premise was that cheaper, more advanced AI will make confidential local or end-to-end protected use increasingly important. But the only reply challenged the claim at its foundation, asking which concrete protection a TEE would improve. There was no resolution: commenters agreed implicitly that model quality alone is insufficient, while splitting over whether the main weakness lies in Claude’s app, the rival product’s capabilities, or the proposed privacy stack.
Where opinion split
The sharpest dispute is whether the competing assistant actually has a better product implementation than Claude. One side treats it as more polished outside the model itself; the rebuttal says that polish is superficial because weak memory and poor generated artifacts make it unable to build useful things.
Coverage & discussions
Hacker News · claude.com
Community discussions
Community Sentiment
Positives
- Claude’s underlying model still commands clear confidence, making application reliability—not baseline intelligence—the perceived risk to its usefulness.
- Trusted-execution environments are seen as a promising route to more private AI use as capable models become cheap enough to run in more settings.
- The rival assistant’s search and information aggregation are acknowledged as genuinely useful, even by a commenter otherwise harsh on its capabilities.
Concerns
- A buggy Claude Duet application could erase the practical benefit of a model users otherwise rate highly, because dependable interaction is part of the AI product.
- The rival assistant is criticized for failing at persistent memory and substantive creation, leaving its outputs too flimsy for tasks beyond retrieval or shopping help.
- The privacy proposal rests on an asserted TEE benefit that remains unsubstantiated in the thread, with no concrete threat model or protection explained.
