What happened?
Anthropic announced Claude Duet, a new proactive AI agent app built around the Claude Opus 5.5 model. Available as a research preview for Pro and Max subscribers in the US, Claude Duet works as an assistant that monitors connected apps, sends morning briefings, and executes tasks in the background.
Unlike traditional chats, Claude Duet takes the initiative: it prioritizes what needs attention, completes requested tasks, and only interrupts the user when a decision is needed. The app is available on iOS, Android, and via claude.ai.
Why does this matter?
Claude Duet represents an important step in the evolution of autonomous AI agents, competing directly with products like Meta's Muse and OpenAI's Dots. It integrates over 6,000 connectors from the Claude Marketplace, including Gmail, Slack, QuickBooks, HubSpot, and Shopify, enabling complete workflows for professionals and small business owners.
Anthropic emphasizes safety features: the agent requires explicit approval before sending messages, making payments, or signing documents, and has dedicated classifiers to protect sensitive information such as credit card numbers.
What changes in practice?
Users can receive a daily summary of priorities, ask Claude Duet to build dashboards, financial forecasts, prototypes, or even web apps, all while it continues working in the background on a virtual machine with its own browser. The preview uses the same usage limits as Pro and Max plans.
The company warns that this is an early version that may still misjudge urgency or misread contexts, and is less strong at personal tasks like booking tickets. Access is via waitlist for eligible users in the United States.
By GeekikiBot
