Claude Duet in the UK: What the Waitlist Does and What It Cannot Do

"Claude duet uk" is a Breakout query, and the honest answer to what it is asking is short: Anthropic has published no UK availability, no UK date, and no waitlist that promises either. The product page describes Claude Duet as a research preview on a waitlist for Claude Pro and Max subscribers in the United States, powered by Claude Opus 5.5. OrcaRouter hosts that model, not the Duet app. We serve Claude Opus 5.5 through an API, and we have nothing to do with the preview, the waitlist or the standalone app, so this article keeps those two things apart from the first sentence onward.

That separation is the whole answer for a reader outside the United States, and it is worth spelling out because most of the search interest in the product is coming from outside the country where it runs. There are two questions tangled together. One is "when can I use the app", and the only published answer is a waitlist with conditions. The other is "can I use the model", and that one has an answer today that has nothing to do with a queue.

What the waitlist is, and what it is not

The page's call to action is "Join the waitlist", and its FAQ supplies the definition: "Claude Duet is a research preview on a waitlist for Claude Pro and Max subscribers in the US." That sentence is the entire published eligibility rule. A plan, a country, and a queue, in that order.

What the page does not do is describe the queue. There is no position, no queue length, no figure for how many people have been invited, and no statement about how long anyone waits. Everything after the submission sits outside what the page documents. A reader looking for a number to watch will not find one, because the number has not been published.

The waitlist is also not a sign-up for the product in the ordinary sense. Claude Duet exists as a standalone app for iOS and Android and on the web at claude.ai on a computer, and the page says so plainly. But access to it runs through the preview rather than through a download. You cannot install your way in.

The plan condition, and why upgrading does not clear it

Eligibility has two conditions, and they are joined. A Pro or Max subscription is necessary but not sufficient; the subscription also has to be attached to a subscriber in the United States. That conjunction is why a plan change on its own moves nobody.

A personal upgrade from Pro to Max satisfies the plan half and leaves the country half exactly where it was. The other direction is worse: Team and Enterprise plans are named on the page as "not included yet", so a workspace moving from Pro to Team is moving away from the published eligibility rather than toward it. The page names Team and Enterprise as the one category it excludes, which makes a business upgrade the one plan change that is clearly not a route in.

The prices are worth stating because they are often mixed up with Duet itself. Anthropic lists Pro at $17 a month billed annually, or $200 up front for the year, or $20 month to month. Max comes in two tiers, 5x at $100 a month and 20x at $200 a month. Team is $20 per seat per month for 2 to 150 seats. None of those is a Duet price. During the preview, Duet is included in the plan rather than sold separately, so paying more for a bigger plan buys usage headroom, and the page publishes nothing that ties plan size to queue priority.

The country condition, and what a VPN does not do

The page does not describe how the country is established. It does not say whether the waitlist form rejects a submission from outside the United States, what checks run on an account, or whether a plan's billing country is the thing that counts. All of that is unpublished, and it is unpublished in the specific sense that matters here: there is no stated mechanism for a reader to reason from.

What is published is narrower. The preview is for subscribers in the US. The app runs on iOS and Android and on the web at claude.ai. Nothing on the page describes changing the country attached to an account, and nothing on the page describes a way to move up or around the queue.

This is not a workaround guide, because no workaround has been published. The page does not mention network location at all. A reader who assumes a VPN changes eligibility is assuming a mechanism the vendor has never described. And even in the case where a form accepted a submission, the waitlist it feeds is defined as a United States preview, so the submission would not change the definition it is being measured against. The constraint is in the sentence, not in the form.

The searches outside the US are louder than the ones inside it

Google Trends, worldwide, past seven days, read on 23 October 2026, makes the mismatch visible. The seed "claude duet" returns 19 rising queries, all of them marked Breakout, the label Trends applies to its fastest-rising terms. The list includes "claude duet uk" and "claude duet sign up" alongside "claude duet release", three different ways of asking the same question in the same week.

The regional split is more pointed still. China leads at 100 on the Trends index, followed by Israel at 70, Singapore at 55, the United Arab Emirates at 53 and the United States at 38, across 66 regions with low-volume regions included. The country where the preview is actually running ranks fifth for interest in it. Whatever the waitlist is doing, it is not absorbing the attention.

One caution changes how far those numbers can be pushed. Trends also surfaces a rising topic called "Claude Paul Duet", a person whose name collides with the product query. Some of the volume is not about the product at all. The figures describe search interest in the query, not demand for the app, and they should not be read as a count of people who want access.

The direction of the gap is still fair to describe. A preview limited to one country is being searched for in dozens of them, and "claude duet uk" and "claude duet sign up" are among the fastest-rising terms in the set.

What you can use outside the US today

The model behind Claude Duet is Claude Opus 5.5, and model access is a separate product from the app. This is the part of the story that does not wait for a country list, and it is also the part we are responsible for.

OrcaRouter hosts Claude Opus 5.5 at $4.00 per million input tokens and $20.00 per million output tokens, with a 1M-token context window and cache reads at $0.20 per million. It sits behind one OpenAI-compatible endpoint alongside more than 200 models, on one key, with automatic failover across providers. We pass provider list prices through at zero markup, so a vendor price change is live on our side the same day. A developer account and a key are what that door requires, rather than a waitlist and a postal address.

What that door does not lead to matters just as much. An API call gives you a model that reasons through a task. It does not give you the morning brief, the flags in your inbox and calendar, the drafted replies waiting for approval, the payroll review, the contract preparation, or the 6,000-plus connectors and plugins. Those are Claude Duet features. We host none of them, and no configuration on our side turns a chat completion into the app.

The distinction sounds obvious written down, and it is routinely muddled in practice. "Powered by Opus 5.5" is a statement about what runs inside Duet, not a statement that Duet is the model. The app is Anthropic's product, sold inside Anthropic's plans, gated by Anthropic's country list. The model is a model, and models are reachable by other routes.

What "sign up" actually gets you

"Claude duet sign up" is Breakout too, and the page offers exactly one sign-up: the waitlist form behind the Join the waitlist button. It does not describe a Duet account, a Duet subscription, or a separate price for the product. The only account in the picture is the Claude account the app runs on.

That has a practical consequence for a reader outside the United States. Joining the waitlist is a statement of interest in a preview whose published audience is US Pro and Max subscribers. It is not a reservation, and the page publishes no position to hold. If you are outside the country, the submission does not make you a preview user today, and that is not a guess about the queue; it is the sentence that defines the queue.

It also means there is nothing to buy. Because Duet is included in the plan during the preview rather than sold separately, a reader who expects a checkout page after signing up is looking for a product Anthropic has not put on sale.

What would actually change the answer

Very little of this is ambiguous, which is the useful part. Three sentences on the product page carry the whole rollout, and each one would have to change for a reader outside the United States to get a different answer.

The first names the country: "in the US". The second names the plans: "Pro and Max subscribers". The third carries the exclusion: "Team and Enterprise plans are not included yet". A UK date would appear as an edit to the first. Business-tier inclusion would appear as an edit to the third. An end to the preview would replace the words "research preview" with something else. As of the page read on 23 October 2026, none of those has happened, and no date for any of them has been published.

That is a more useful thing to watch than a search query. Queries rise and fall on rumour; the eligibility sentence changes only when the product does.

The two doors, side by side

A reader in the UK has, in practice, two separate questions and two separate answers. The app is unavailable, undated, and gated by a waitlist whose queue is not published. The model is available through an API today, priced per token, through a developer account rather than a waitlist.

Confusing the two produces the worst of both: waiting for an app because the model sounded close, or buying tokens expecting a morning brief. They are not substitutes. One is a workflow product that watches your inbox and asks for approval; the other is a reasoning engine you call from your own code. Anthropic has been unusually clear about which is which. The search data suggests most of the world is still looking for the answer in the wrong place.