Anthropic opens a research preview of Claude Duet, a proactive agent app that asks before acting

Anthropic on 21 October 2026 opened a waitlisted research preview of Claude Duet, a separate app in which Claude acts as a proactive agent: instead of waiting for a prompt, it comes to the user with what needs attention.

What it does

• Sends a short morning brief on the most important items across connected apps such as email and calendar, and checks those accounts every few hours during the day.
• Takes tasks by text or voice message, from scheduling meetings to building dashboards, forecasts or simple web apps, and works on them in a cloud virtual machine with its own browser, so the user’s phone or laptop can be off.
• Reuses the memory, skills and connectors from the user’s Claude account. Anthropic says it is built around its Claude Opus 5.5 model.

Guardrails

Before Claude Duet sends a message, makes a payment, signs a document or deletes anything, it shows the exact action and waits for approval. Anthropic says some actions, such as trading securities or moving money between accounts, are off-limits even with approval. The company also warns that the early version may misread a thread or misjudge what is important.

Availability

The preview is available on iOS, Android and claude.ai to US users on Pro and Max plans via a waitlist and uses the same usage limits as the rest of Claude. Team and Enterprise plans are not included yet. CNBC framed the launch as Anthropic’s answer to personal agents from rivals such as Meta and OpenAI.