Claude Age Verification: Why You're Locked Out

Claude age verification is the check Anthropic runs when its systems flag an account as possibly belonging to someone under 18: the account gets disabled, and a notification email carries a link to Yoti, a third-party age verification service. Pass the check and the account comes back. The rule itself is blunt, and Anthropic's help centre states it in one line: "Claude, our consumer product, is only available to people over 18 years."

Key Takeaways:

• Anthropic disables consumer Claude accounts "based on indicators of minor activity", then offers reinstatement through Yoti
• Three ways to pass: a selfie for facial age estimation, a photo of an ID document, or an existing Yoti Digital ID app attribute
• Anthropic says it receives only a pass/fail result and never sees your ID or your photo; Yoti deletes the images once the check completes
• The policy page is dated 19 May 2026, so this is not a rule that appeared overnight; what changed on 11 September is that it reached the front page of Hacker News with 572 points
• The rule covers the consumer product. Anthropic's help centre does not describe the same gate on API or Console access

What Claude age verification actually is

Anthropic's support article, Age assurance on Claude, describes a two-stage system rather than a one-time form. Stage one is the checkbox during signup, where you confirm you're 18 or over. Stage two runs quietly afterwards: "We have safety systems in place to detect if people under 18 may be using Claude and we'll disable accounts based on indicators of minor activity."

That second sentence is the one worth reading twice. The account isn't put into a limited mode or asked to verify before continuing; it's disabled, and the invitation to verify arrives afterwards by email. So the first sign most people get is that Claude has stopped working.

What trips the check, and what a disabled account looks like

Anthropic doesn't publish the signals that trigger Claude age verification. The help centre says only "indicators of minor activity", which leaves the detection logic deliberately unspecified, and there's no public list of what counts. Anyone telling you exactly which prompts flag an account is guessing.

What users describe publicly is more useful than speculation about the model. Several people posting on X in 2026 reported suspensions carrying a reason along the lines of "account used by a child", including one who said they had already passed the Yoti check once and were flagged again afterwards. Another described being told they "prompted like a child". Those are individual reports rather than Anthropic's own wording, and they're worth reading as a pattern rather than as documentation: the classifier is looking at how you write, not at a birthdate you typed once.

The practical consequence of a false positive is the part people underestimate. Your projects, your chat history, your connectors and any scheduled work attached to that login all go quiet at the same moment, and the route back runs through a third-party identity check rather than through support. If you've read our write-up on an Antigravity account suspended over a third-party integration, the shape will be familiar; the trigger differs, the blast radius doesn't.

The first hour after the account goes dark

Order of operations matters here, because two of the obvious instincts waste time. Clearing cookies, switching browsers and creating a fresh account with the same email all do nothing to a decision taken at the account level, and the last one risks compounding the problem rather than solving it.

Start in the inbox instead. The reinstatement link arrives by email, and the help centre is explicit that "a link for verification will be provided in the notification email", so that message is the only route the documentation describes. Check the spam folder before concluding it never arrived; the sender is a transactional address rather than the one your existing Claude mail comes from.

Then look at what the email actually says before you upload anything. Claude age verification is one of several reasons an account can be disabled, and an email with no Yoti link is telling you the trigger was something else, which means the fix is a support conversation rather than a selfie. A handful of the X posts from this year describe exactly that confusion: people hunting for an age verification flow when the notice they received pointed at the usage policy instead.

If the link is there, have the document ready before you open it. Session timeouts during a half-finished identity check are a common way to end up starting over.

The three Yoti methods, and the documents that work

Anthropic hands the actual check to Yoti, and the help centre lists three routes, any of which you can pick:

• Facial age estimation. Take a selfie; Yoti's model estimates your age from the image. No document required, which makes this the fastest path for anyone who doesn't want to photograph a passport.
• ID verification. Photograph an identity document and upload it from your device.
• Yoti Digital ID app. If you already use the app, share a verified "over 18" attribute from it instead of starting fresh.

The document route is narrower than "any government ID", and the list is worth checking before you start. Yoti's supported documents page covers passports universally, plus national ID cards, driving licences and residence permits. Beyond that it gets country-specific: CitizenCard, the Post Office card and the Young Scot National Entitlement Card in the UK; provincial and territorial ID, health cards and the Secure Certificate of Indian Status in Canada; State ID in the US; NEXUS on both sides of that border; Aadhaar and PAN in India; MyKad in Malaysia.

If your only photo ID is a student card or a library card, none of the three routes will take it, and facial age estimation becomes the realistic option rather than a preference.

What Anthropic sees, and what Yoti keeps

This is the part that reads better than the headline suggests. Anthropic's page is specific about the data path: "Anthropic never sees your ID or image; we receive only a pass/fail result and do not process or store any personal data from the verification." Yoti's side is described as equally short-lived, with your selfie and document images "deleted by Yoti as soon as your age is checked", and Yoti itself described as independently audited for SOC2 compliance.

Two boundaries define the data path in Claude age verification. Your document goes to the verification provider and not to the model company, and the result that comes back to Anthropic is one bit wide. That's a meaningfully better arrangement than uploading a passport into the same account that holds your chat history, and it's the design most age-assurance deployments have converged on since the UK and several US states started requiring checks on other categories of site.

It still means handing a photograph of a passport to a company you had no relationship with an hour earlier. Reasonable people land on both sides of that trade, and the Hacker News thread is mostly an argument about exactly this. Anthropic's page doesn't claim otherwise; it describes the handling, not the comfort.

Why 11 September looked like the announcement, and wasn't

The support article carries a date of 19 May 2026. Reports of age-related suspensions and Yoti prompts go back further still, with users posting about Claude account bans over suspected minor status across the northern spring and summer. What happened on 11 September 2026 is that someone submitted the help centre URL to Hacker News under the title "Claude is only available to people over 18 years", and it climbed to 572 points and 596 comments.

That distinction matters if you're trying to work out whether your own account is at risk today. Nothing in the policy changed this week; the audience did. An enforcement system that had been running quietly for months became a topic, which is also why the thread is full of people discovering a rule that had already been applied to them.

Keeping recurring work off a single consumer login

There's a boring operational lesson underneath the privacy argument, and it has nothing to do with age. Any consumer account can be disabled by an automated review, and when it is, everything hanging off that login stops at once. Compliance is one trigger. Billing failures, fraud heuristics and regional restrictions are others, and the response time is the same in every case: however long a human or a verification provider takes.

Work that runs on a schedule is the part that suffers most, because nobody is watching it fail. A weekly report that quietly doesn't generate costs you more than a chat window you can't open, since you find out about the chat window immediately. Zentor sits in a different place in that stack: it's a hosted cloud AI computer with its own machine and its own account, so a job that runs there isn't tied to the session state of a desktop chat app. Running it alongside your existing Claude subscription is the point; it isn't a replacement for the assistant you already like, and it is emphatically not a way around an age rule, which applies wherever you go.

The cross-device property is the second half of it. Because the environment lives in the cloud rather than on the laptop you happened to be signed in on, picking the work back up from a different machine doesn't involve re-establishing anything locally. Free access at Zentor runs three days on 1,000 credits and the $20 plan renews 1,000 credits monthly, which is a different commitment shape from a single consumer login carrying everything you've automated. For the longer argument about where scheduled agent work should live, our guide to managed AI agent services covers the trade-offs, and Claude API for AI agents covers the case where you'd rather build against the API than the consumer app.

FAQ

Why is my Claude account disabled?

If the disabling came with an email offering age verification through Yoti, Anthropic's systems flagged the account under its minor-activity detection. Other disable reasons exist and carry different notification wording, so the presence or absence of a Yoti link in the email is the fastest way to tell which one you're dealing with.

How do I verify my age on Claude?

Claude age verification runs through the link in the notification email, which opens Yoti. Choose facial age estimation for a selfie-only check, ID verification to photograph a document, or the Yoti Digital ID app if you already hold an "over 18" attribute there. Anthropic says the account is reinstated once you pass.

Does Anthropic see my ID?

No, according to its help centre. Anthropic states it receives only a pass/fail result, and that Yoti deletes your selfie and document images as soon as the age check is finished.

Can under-18s use Claude at all?

Not the consumer product. Anthropic's wording is that Claude is "only available to people over 18 years", and the detection system exists to enforce that rather than to grade it. Anthropic publishes separate guidance for organisations serving minors, which is a different arrangement from an individual consumer account.

Is there a regional difference in how this is enforced?

The help centre doesn't say. It describes one rule for the consumer product without carving out regions, and we found no official statement distinguishing EU or UK enforcement from anywhere else, so treat any claim about regional variation as unverified.

Verified against Anthropic's help centre, Yoti's developer documentation and the Hacker News thread on 12 September 2026. Anthropic does not publish the detection signals, and the user reports quoted here are individual accounts rather than company statements.