Anthropic challenges Grok Bot, Meta's Muse, and OpenAI's Dots with its always-on agent Claude Duet.


Anthropic has introduced a research preview of Claude Duet, a standalone app that answers Grok Bot, Meta's Muse, and OpenAI's Dots with a proactive Claude agent. It matters because Duet sends a short summary of the most important items from connected apps every morning, continues tasks on a virtual machine while the user is away, and asks for approval before sensitive steps.

Competitor Moves Opened the Personal Agent Window

The launch follows a cluster of similar attempts. SpaceXAI opened the Grok Bot beta in August. In September, Meta released Muse and OpenAI introduced Dots. These products all lean toward personal agents that take initiative rather than waiting inside a chat window. Claude Duet enters that race by reversing the usual flow: instead of the user opening a chat, the agent comes to the user. After the morning summary, it reaches out only when something needs attention.

From Morning Summaries to Connected Work

The agent pulls memory, skills, and app connections from the user's existing Claude account. A single ongoing thread accepts tasks by text or voice message, covering scheduling, dashboards, financial forecasts, prototypes, and web apps. Duet connects to Gmail, Slack, Stripe, Docusign, and Shopify through more than 6,000 connectors and plugins from the Claude Marketplace. That setup can track unpaid invoices, draft payment reminders, and get payroll ready. For purchases, it uses Stripe Link with a virtual card, so the user's real card number never gets shared. The agent works on a virtual machine, its own cloud computer with a browser, so the user's laptop or phone does not need to stay on.

Approval Rules, Safety Layers, and Known Gaps

Anthropic says Duet can only read, draft, and prepare on its own. Before it sends a message, makes a payment, signs a document, or deletes anything, it shows exactly what it plans to do and waits for the user's approval. Trading securities and moving money between accounts are off the table, even with approval. Credential handling follows the design of Claude Managed Agents, so the agent never sees passwords or login tokens directly. For prompt injections, hidden instructions that can manipulate an agent, Anthropic uses defenses from its research on browser use. Red teams and an outside firm also ran attack tests to probe Duet for weak spots. A new classifier works with special safety instructions in the system prompt to check every message for leaking sensitive data. Together, they are meant to refuse requests to share credit card numbers or ID details. Health or personal finance information only gets through when the user explicitly says so.

Early Limits Shape the Roadmap

The early version can misread a message thread, misjudge urgency, and miss things that arrive between two checks. It is also weaker at personal errands like booking tickets, since its core features are built for work. Duet checks connected accounts every few hours, and Anthropic says that comes down to usage. Constant checking would burn through subscription limits much faster. Duet shares those limits with other Claude products, and the background checks count against them. Anthropic plans to use tester feedback to tune when the agent should reach out. How Duet fits with the Claude app is still an open question. For now, it runs separately, and early users' experience will shape how the two work together.

Meta description: Anthropic's Claude Duet research preview turns Claude into a proactive agent that summarizes work, checks apps, and asks for approval before sensitive steps.
Tags: Claude Duet, Anthropic, AI agents, prompt injection, Stripe Link