Anthropic has launched Claude Duet, a separate app in which its Claude model works as a background agent. Instead of waiting for you to type a request, Claude Duet checks your connected work apps, flags what needs attention and carries out tasks while you are away, asking for approval before it acts on your behalf.
The company announced it on 21 October as a research preview. It is built around Claude Opus 5.5 and is aimed at professionals and small business owners. Access starts with US subscribers, so readers in Africa and elsewhere will have to wait. Here is what it does, what it costs and where it falls short.
What Claude Duet does
According to Anthropic's announcement, Claude Duet works in one ongoing conversation that you can type or speak into. Its main features are:
- A morning brief. Each morning it summarises the most important items across your connected apps.
- Background checks. It looks at connected accounts roughly every few hours and messages you only when something needs attention.
- Tasks of different sizes. From scheduling a meeting to building dashboards, sales or financial forecasts, prototypes and small web apps.
- Its own computer. It runs on a virtual machine with its own browser, so your phone or laptop does not need to stay on. You get a notification when a task is done or needs review.
- Your existing setup. It inherits the memory, skills and connected apps from your Claude account.
Examples Anthropic gives
Claude Duet can use more than 6,000 connectors and plugins from the Claude Marketplace. Anthropic's examples include finding missed Gmail and Slack messages and drafting replies, checking QuickBooks or Stripe for unpaid invoices and drafting reminders, preparing payroll in Gusto, sending agreements for signature through Docusign, drafting follow-ups from HubSpot deal history, and setting up a Shopify store.
It can also buy things. It uses Link, Stripe's digital wallet: Claude Duet confirms the item with you, you approve the spend in Link, and it pays with a virtual card so your real card number is not shared. It can also check out with Shop Pay.
Testers quoted by Anthropic include a property broker who got an unprompted flag about a showing booked on the wrong day, and a restaurant owner with about 110 employees who uses it for accounting, support and store-by-store invoices.
Claude Duet pricing and availability
- Price: no separate fee. It is included in Claude Pro and Max plans.
- Usage limits: shared with the rest of Claude. Background checks count against them, and big jobs such as financial models use more.
- Who gets it first: US Pro and Max subscribers, through a waitlist at claude.com/product/claude-duet.
- Platforms: standalone iOS and Android apps, plus claude.ai on the web.
The announcement mentions only the United States. Anthropic has not said when other countries, including African markets, will get access. Many of the connectors it highlights, such as Gusto payroll, are also built for US businesses.
Safety and permission controls
An agent that can read your email and spend money needs guardrails. Anthropic lists several:
- Approval first. Before it sends a message, makes a payment, signs a document or deletes anything, it shows the exact action and waits for you to approve.
- Hard limits. Trading securities and moving money between accounts are blocked even if you approve.
- No direct access to passwords. Credentials are handled so the agent never directly sees passwords or sign-in tokens.
- Sensitive data. A dedicated classifier blocks sharing of the most sensitive items, such as card numbers and government ID. Health and personal finance information is shared only when you explicitly ask.
- Checks on actions. Actions pass through Anthropic's auto mode classifier, with defences against prompt injection drawn from its browser-use research. The product has had red teaming and an external penetration test, and is covered by Anthropic's bug bounty.
Anthropic says it will publish more on the safety design "in the coming days."
Limits Anthropic admits
The company is open that this is an early version. It may misread a thread or misjudge how urgent something is. Personal tasks such as booking tickets are weaker than work tasks. You cannot yet watch its browser. It lives in its own app, separate from the main Claude app, and Anthropic is still tuning when it decides to message you.
Should you care if you are outside the US?
Not yet as a user, but the direction matters. Claude Duet is part of a shift from chatbots you prompt to agents that watch your accounts and act. That raises practical questions about privacy and control, the same ones we looked at in our explainer on ChatGPT's locked chats.
For small businesses in Lagos, Nairobi or Douala, the tools that would make an agent like this useful, from accounting software to payment wallets, vary by country. When access widens, the real test will be whether it connects to the apps African businesses actually use.
Source: Anthropic, "Introducing Claude Duet", 21 October 2026.
