Anthropic Launches Claude Duet Agent for Pro and Max Subscribers
Anthropic introduces Claude Duet, a proactive AI agent that manages tasks across connected apps and runs in the background.
Key points
- Claude Duet is rolling out as a research preview to US Pro and Max subscribers.
- The agent runs on a cloud computer, allowing work to continue after the app closes.
- Anthropic states the feature counts toward existing plan usage limits at no extra cost.
Anthropic announced Claude Duet today, a proactive AI agent designed to manage tasks across connected applications. The company is currently rolling out the feature as a research preview to Pro and Max subscribers in the United States. Access is expanding gradually through a waitlist system.
The agent operates on a dedicated cloud-based computer, enabling it to continue working even after users close the application. Anthropic notes that Claude Duet uses the same memories, skills, and connectors as the standard Claude model. It can search past conversations to avoid requiring users to repeat information.
Context and Background
Anthropic describes Claude Duet as a competitor to Meta’s Muse and OpenAI’s Dots, targeting professionals and small business owners. The agent provides morning briefs by scanning email and calendar data to identify items needing attention. It drafts replies for review and follows up on inactive threads throughout the day.
The system integrates with Stripe to check for unpaid invoices and draft payment reminders. It also connects with Link and Shop Pay to execute purchases, but only with explicit user approval. Anthropic emphasizes that the agent asks for permission before sending messages, making payments, signing documents, or deleting files.
Who Is Affected
Claude Duet is currently unavailable for Team or Enterprise plans, though Anthropic plans to expand availability to those subscriptions in the future. The feature is included at no additional cost for eligible subscribers. However, background work counts toward existing plan usage limits, which may cause users to reach those caps sooner than with chat alone.
The agent is available as a standalone app on iOS and Android, as well as through Claude on the web. It is built around the Claude Opus 5.5 model. Anthropic warns that this early version may misread threads or miscalibrate urgency, and it is not yet strong at personal errands like booking tickets.
What should you do?
Users should monitor their plan usage limits closely, as background agent activity consumes credits. Anthropic states that Claude Duet includes a dedicated classifier and safety system prompt designed to reject requests to share sensitive information, such as credit card numbers. The company advises treating the preview as an early version with potential limitations.
What to do and how to stay safe: Anthropic
- Review your current subscription usage limits before enabling background agent tasks to avoid unexpected caps.
- Verify that all connected applications have appropriate permission scopes and that you approve each significant action.
- Monitor the agent’s morning briefs and daily updates to ensure it correctly interprets urgency and context.
- Avoid sharing strict-tier sensitive information like credit card numbers, as the safety system is designed to reject such requests.
Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
Is Claude Duet available for Enterprise plans?
No, Anthropic states that Claude Duet is not available on Team or Enterprise plans yet, but plans to expand availability.
Does Claude Duet cost extra?
No, the feature is included at no additional cost, but it counts toward existing plan usage limits.
Can Claude Duet make purchases?
Yes, it integrates with Link and Shop Pay to make purchases, but only with explicit user approval.
