Anthropic has introduced Claude Duet, a new AI agent aimed at professionals and small business owners. The company is positioning it as a response to other agent-style products such as Meta Muse and OpenAI Dots, and it is being released as a research preview. That framing places Claude Duet in a fast-moving category where AI products are increasingly expected to do more than answer questions: they are being designed to help people manage ongoing work, keep track of information, and take on tasks across multiple connected services.
A new agent for connected work
According to the announcement, Claude Duet is designed to work across connected apps and help people stay on top of what matters during the day. Anthropic says the agent can send a short brief each morning covering the most important items across those apps, then continue monitoring them and notify the user only when something needs attention. In other words, the product is meant to function less like a one-off assistant and more like a persistent layer between the user and the apps they already rely on.
The product is available as a standalone app on iOS and Android, and it can also be used through Claude on the web. Anthropic says users can hand it tasks, making the agent part assistant and part daily organizer for people who work across multiple services. That combination of app access, task support, and continuous monitoring is central to how Anthropic is describing the new product in its preview stage.
For professionals and small business owners, this setup suggests a workflow-focused assistant rather than a general-purpose chatbot. The emphasis on connected apps means the experience is tied to information already flowing through the user’s services, while the morning brief and ongoing alerts are meant to reduce the need to check everything manually throughout the day. Anthropic is presenting Claude Duet as a tool that can keep pace with work as it happens.
Built on Claude Opus 5.5
The Verge reports that Claude Duet is built around Claude Opus 5.5. That detail places the new agent within Anthropic’s broader Claude lineup while giving the company a named model foundation for the product’s preview launch. It also helps define Claude Duet as part of a larger family of products rather than a standalone experiment detached from the rest of Anthropic’s work.
Engadget reports that Anthropic subscribers in the United States can join a waitlist to try the new agent. Both reports frame the release as an early look rather than a full general rollout. The preview status matters because it signals that Anthropic is still testing how Claude Duet performs in practical settings and how users respond to its mix of briefing, monitoring, and task delegation.
With a research preview, Anthropic is able to introduce the agent while keeping its rollout limited. That gives the company room to observe how people use the product across iOS, Android, and the web, and how well it handles the connected-app workflow it is designed around. The waitlist format also reinforces that the launch is intentionally controlled, with access not yet open in a full public release.
What the preview suggests
For now, Claude Duet appears aimed at users who want a single agent to summarize activity, watch for important changes, and accept delegated tasks across the apps they already use. The combination of a morning briefing, ongoing monitoring, and task handling suggests Anthropic is trying to make the assistant useful throughout an entire workday instead of limiting it to single prompts or isolated actions. That broader role is what makes the product feel closer to a workflow companion than a conventional chat interface.
The launch also shows how quickly AI agents are becoming a competitive category. Anthropic is entering a space where agent products are now being framed not just as chatbots, but as tools that can sit between users and their apps, filter information, and help coordinate routine work. In that sense, Claude Duet is part of a wider push toward agents that can stay active in the background and react when something matters, rather than waiting for every instruction to come from the user.
The comparison to Meta Muse and OpenAI Dots makes the competitive context especially clear. Anthropic is not presenting Claude Duet as a small feature update, but as an answer to a growing category of products built around the idea of persistent digital assistance. The research preview gives the company a chance to test that idea in a real product while keeping the launch aligned with its current state of development.
Conclusion
Claude Duet is Anthropic’s latest step into the AI agent market, and its research preview status means the company is still testing how the product fits into real-world workflows. With support for iOS, Android, and the web, plus a focus on connected apps and task support, the release gives Anthropic a direct answer to the emerging agent products from Meta and OpenAI. For now, the launch reads as both an early product test and a clear statement of where Anthropic wants to compete next.