Claude Duet (Anthropic)
claude.com/product/claude-duet ↗Claude Duet in short
What it is: Best for US Claude Pro, Max or Ultra subscribers who want morning briefs and check-ins; it is a waitlisted research preview that shares your Claude limits. Price: Paid. No separate price: part of a Claude Pro plan ($17/mo billed annually or $20 monthly) Max ($100/mo) or Ultra ($200/mo), US only, by waitlist. Uses the same usage limits as the rest of Claude. (checked 2026-11-03). Best for: get a morning brief of what needs your attention; let an agent watch your connected accounts and flag what matters. Watch out: research preview with a waitlist, for Claude Pro, Max and Ultra subscribers in the US only.
Our verdict
Best for US Claude Pro, Max or Ultra subscribers who want morning briefs and check-ins; it is a waitlisted research preview that shares your Claude limits.
Anthropic's product page describes Duet as a proactive agent in research preview that works on its own cloud computer, asks before sensitive steps and draws on your Claude plan's limits.
Based on public docs and pricing pages, not hands-on testing. How we checkStrengths
- Proactive: checks your connected accounts and messages you when something needs you
- Works on its own computer in the cloud using the apps you connect
- Asks before sensitive steps such as sending messages or paying
- Standalone app for iOS and Android, plus claude.ai on a computer
Limitations
- Research preview with a waitlist, for Claude Pro, Max and Ultra subscribers in the US only
- Team and Enterprise plans are not included yet
- Shares your Claude plan's usage limits, and background checks count toward them
- Anthropic says it may misread a thread or misjudge what is urgent
Claude Duet pricing
No separate price: part of a Claude Pro plan ($17/mo billed annually or $20 monthly) Max ($100/mo) or Ultra ($200/mo), US only, by waitlist. Uses the same usage limits as the rest of Claude.
From the official pricing page, checked 2026-11-03. Plans change, so confirm before you buy.
Is Claude Duet free?
Not on a general free plan, per its pricing page. No separate price: part of a Claude Pro plan ($17/mo billed annually or $20 monthly) Max ($100/mo) or Ultra ($200/mo), US only, by waitlist. Uses the same usage limits as the rest of Claude. (official pricing page, checked 2026-11-03)
Good for
Note: Separate from the main Claude app and from Claude Tag (a Slack teammate for Team and Enterprise). Announcement, Oct 21, 2026: https://claude.com/resources/articles/claude-duet
Compare all AI assistants: pricing, credits, approvals and data use →
What users say
Community sentiment (opinion, not verified facts)Claude Duet launched Oct 21, 2026, a day before we checked, so there is very little community discussion yet. Early hands-on reviews are in the guide's expert section.
Interest in the waitlist, and US only access
Early posts are mostly people joining or asking about the US waitlist, plus launch news.
- @CampbellKaleb23, Oct 21, 2026 on X
- @PelletApp, Oct 21, 2026 on X
Full community section and how we gathered it, in the AI assistants guide →
Alternatives in Assistants
- Muse (Meta): Best for trying a personal agent for free, with approvals and an audit trail; it is offered in North America, Japan and, since Oct 27, 2026, Singapore.
- Dots (OpenAI): Best for ChatGPT Pro users who want always on agents in ChatGPT, Slack and Teams; not offered in the EEA, Switzerland or the UK.
- Grok Bot: Best for running several agents that share a cloud computer, tools and files; the weekly pool is shared with Grok and Cursor.
- ChatGPT Work (OpenAI): Best for turning your tools and files into documents, presentations and web apps; the Free plan is desktop only, and usage is shared with Codex.
- Claude (Anthropic): Best for work on your own files and browser when you want to choose between approving each action and automated checks; local files and the browser need Claude Desktop open.
- Manus: Best for trying long tasks on free daily credits; monthly credits reset and go to zero if you cancel.
- Lindy: Best for teams that want an assistant in Slack and email that waits for approval on outside actions; every active user is billed.
- Motion: Best for planning your week with calendar scheduling; trials bill unless you cancel.
- Alexa+ (Amazon): Best for Prime households that want errands like takeout, reservations and rides done by voice; check approvals yourself, since Amazon's pages do not describe them.
- Comet (Perplexity): Best for an assistant inside a free browser; agent work in Work mode runs on paid credits.
- Copilot Cowork (Microsoft): Best for Microsoft 365 Premium or Pro subscribers; Personal and Family plans lose access, and it is still a preview.
- Gemini (Google): Best for Google AI Pro or Ultra subscribers who want background tasks with apps like Instacart and OpenTable; not offered in the EEA, Nigeria, Switzerland or the UK.
Recent Claude Duet releases
- Claude Duet (research preview) (2026-10-21)
FAQ
Is Claude Duet free?
Not on a general free plan, per its pricing page. No separate price: part of a Claude Pro plan ($17/mo billed annually or $20 monthly) Max ($100/mo) or Ultra ($200/mo), US only, by waitlist. Uses the same usage limits as the rest of Claude. (official pricing page, checked 2026-11-03)
How much does Claude Duet cost?
No separate price: part of a Claude Pro plan ($17/mo billed annually or $20 monthly) Max ($100/mo) or Ultra ($200/mo), US only, by waitlist. Uses the same usage limits as the rest of Claude. (from claude.com/product/claude-duet, checked 2026-11-03).
What is Claude Duet used for?
Listed use cases: get a morning brief of what needs your attention; let an agent watch your connected accounts and flag what matters; hand off tasks that keep running while you are away; approve drafted replies before they are sent.
What are the best alternatives to Claude Duet?
Other AI assistants in this directory: Muse (Meta), Dots (OpenAI), Grok Bot, ChatGPT Work (OpenAI). We list them in data order, not ranked.
Did you test Claude Duet?
No. Verdicts are based on each vendor's public docs and pricing pages, not hands-on testing, unless a page says otherwise.