Unclaimed by AI, the name checker: What actually happens when you hit "check"?

A naming studio for the age of AI. One place, from a blank page to a claimed name.

The old version of this problem had two tools: a generator that spit out names nobody would use, and a WHOIS lookup that told you half of them were taken. Neither knew what the other was for. This tool runs the whole path in one place.

• Generate. describe what you are building and a sandbox fills with name ideas, four studios arguing from different angles and a judge keeping the strongest
• Check. turn an idea into names, then check the shortlist live: domains by RDAP, handles on five platforms, and what three AI models already believe the name means
• Watch. star the names you love and Pro rechecks them every few hours, emailing you the morning something gets claimed

Already have a name in mind? Skip the generator entirely. The Name mode runs the same three checks on one exact name and hands you a shareable results page.

Three checks, run together, in front of you. Not a score pulled out of a black box.

• Domains. registered or not on .com, .ai, .io and .co, answered live by the registry itself
• Social handles. free or taken on GitHub, X, Instagram, TikTok and YouTube
• AI associations. what ChatGPT, Claude and Perplexity already believe the name means, asked independently, with no hint about your idea

Every result ships with the evidence behind it, because the point is not a score. The point is that you can judge the risk yourself.

We ask the registry itself, at the moment you click. No cached copy, no WHOIS mirror that might be three months stale.

RDAP is the protocol that replaced WHOIS for registration data, and it is what registrars use internally. If we say a domain is open, it was open when we asked.

What we cannot do is hold it for you. Someone can register the name between our answer and your purchase, so when you settle on a name, buy the domain soon.

It depends on the platform, and we label which is which rather than pretending all five answers are equal.

The big platforms publish nothing that gives a definite answer, so we load the public profile page and reason from what comes back. A best guess is usually right.

Where an exact answer does not exist, we say so on the card instead of faking a green checkmark. Confirm the handle on the platform itself before you print it on anything.

This is the check no domain search will ever catch, and it is the reason this site exists.

900 million people ask ChatGPT something every week. When someone asks about your category, the assistant answers from what it already associates with names in that space. If your name is already tied to a different business, the model will describe that business when asked about yours. Your reviews, your pricing, your reputation, all credited to someone else.

So we ask the models directly. Three of them, separately, with no hint about your idea, so the answer cannot be coached. A domain search or trademark search will never surface this overlap, because it is not a legal question. It is a question about what a machine already believes.

Treat it as an early warning system, not a verdict. And read it asymmetrically.

• Multiple models recognize the name. strong signal. That is a real brand with real presence, and you should probably pick a different name
• No model recognizes it. weak signal. It means they have not come across the name yet, which is not the same as nobody owning it

Models get updated regularly, so a name that looks clear today could be linked to something else later. That is another reason the watchlist exists.

No, and anyone who tells you otherwise is selling something.

• What we check. whether the name is practical to use online: domains, handles, and what the AI models already associate with it
• What a trademark search checks. whether someone holds legal rights to the name, in a specific jurisdiction and class of commerce

Run a name here first. It is a fast way to rule out the obvious conflicts. Then the names you are serious about go through a trademark search or an attorney, because that is a different question with different consequences.

Each extension is its own registry, so availability is independent per extension. There is no single global yes or no.

A name can sit registered on .com for a decade and still be free on .ai, .io and .co, the extensions startups actually reach for. We show all four side by side so you can decide whether another extension suits the brand or whether the taken .com is a reason to pick a different name.

We don't sell domains, so no. What we do is tell you it's open, then hand you a registrar that does.

A domain check and a domain purchase are two different businesses. Ours is the check: live RDAP lookups, five platforms of handles, three AI models, all in front of you with the receipts. The purchase is a registrar's job, and it should stay that way.

So when a name comes back open, we hand you a link to Namecheap, GoDaddy, or Dynadot, and you finish there, in your own name, at their price. Some of those links earn us a commission at no extra cost to you, and we label them when they appear.

We say this plainly because the alternative is a middleman who marks up a domain and calls it convenience. The check is the product. The registration is yours to finish, and it should be in your name, not ours.

Before you grieve, look at how it is taken. Not every conflict is equal.

• Active company in your category. confusion with customers and search results. A different name is the safer choice
• Only one extension or one handle conflicts. a variation or another extension may be enough, and each variation goes through the same three checks in under a minute
• Sitting dormant with no use. a legal question, not a technical one. This tool will not tell you whether you can take it, and no honest tool can

Yes, and you never have to generate anything to do it.

Switch the prompter from Idea to Name, type the exact name, and it runs the same three checks without the generation step, landing on a results page you can share with whoever is arguing with you about this.

Useful when you are comparing a shortlist from a co-founder, deciding between a rebrand and your current name, or testing the name someone suggested at dinner before you spend money on it.

Yes, on Pro. Star a name and it goes on the watchlist, where we recheck it every few hours.

Domains get registered while you sleep. Handles get claimed during your commute. The watchlist is how a name you love stops being a screenshot you forgot and becomes something that emails you the morning it changes.

The streak on each card shows how long a name has stayed untouched, and watching saves you from manually re-running checks while you are still deciding.

A sandbox is a collection of ideas, the way a search is a collection of names.

Describe what you are building, and the sandbox becomes the room where that brief's names get generated, researched and turned into searches. Everything about the project stays in one place instead of scattered across a history list.

Each sandbox carries its own steer, the eight answers about your skills, time, budget and audience that shape generation. A weekend baking idea and a serious invoicing idea never have to share the same steer, and you can override any answer for one sandbox without touching the account default.

Four studios argue from different angles, and a judge keeps the strongest. You never see the argument, only the shortlist.

Why not one model and forty names? Because one model in one mood produces one flavor of name. A list of forty from a single perspective is forty versions of the same idea. Four perspectives produce a range you can actually choose within.

You can also steer a run with a seed, a phrase like "something in fintech", and the studios bend toward it without narrowing to it.

Four sections, bought per idea, written by the same models that generate the names with a research brief.

• Competitors. who else plays here, and how crowded it is
• Demand. whether people are searching for this, and the trend
• Audience channels. where the buyers already are and how to reach them
• Monetization. how the idea earns money, with rough price points

It is one-shot with a paid refresh: a new purchase creates a new version and keeps the old one readable, so the history of what you paid for survives. If research fails to generate, the credits come back automatically. No ticket, no email.

Credits are the pay-as-you-go layer on top of the free plan. Three things cost them, and nothing else does.

They cost $0.50 each at the single rate and less in packs, they never expire, and they only buy those three things. The monthly search allowance on your plan always spends first; credits pick up when it runs dry, so running out of plan never blocks you.

Yes, and the free tier is a real product, not a demo that ends at the good part.

When you need more than that, credits buy research, extra searches and re-checks with no subscription attached. And Pro, at $14.99 a month, is for people who live in the tool: ten sandboxes, fifty searches a month, five checks per name, twenty idea runs a day, and the watchlist.

The distinction is simple. Credits are for occasional depth. Pro is for when naming stops being a one-weekend problem and becomes part of how you work.

No. Your briefs and ideas are yours, and we do not train models on them.

To generate names and research, your text is sent to third-party model providers, because that is where the models live. That is a request, not training data, and the privacy policy states it plainly rather than burying it.

The same honesty applies to the checks: to tell you what AI models associate with a name, the name has to be sent to AI providers. Type brand names here, not passwords.

It stays yours. Nothing is deleted and nothing is held hostage.

Your sandboxes, ideas, searches and check history all survive the downgrade. You drop to the free tier's limits: one sandbox and the monthly free search allowance. The rest of your work is still there, readable.

Buy credits or re-subscribe later and the same account picks up where it left off. Naming is a project you return to, not a treadmill, and the tool should behave that way.