It runs every check on this site against one domain at once, sums them up in a line, and lays the answers out side by side: what is published at the address and how a link to it unfurls on Google, X, Facebook, LinkedIn and iMessage; its certificate and what else that certificate covers; what serves the site and which browser protections it turns on; whether mail is set up and whether anybody can forge it; who holds the name, until when, and the dated path back to the pool if nobody renews; how the zone is wired up; how much of the extension catalogue the word is still free under and which popular extensions are open; who holds the near-spellings of it; what the same name looks like as a handle on twelve platforms; names built from the same word that are free right now; what the name is worth; and how it reads out loud. Each panel is a summary with a link to the tool that answers it in full.

It is the one place on the site that draws a conclusion rather than reporting a reading: registered or free, live or idle or parked, an established brand or an opportunity, and on its way back to the pool or not. Every clause in it comes from a check that actually answered — nothing that timed out or was refused contributes a word — and whatever could not be read is named underneath the sentence, so a summary built from half the page never looks like one built from all of it. It is built from the four checks that land fastest, which is also why it settles in about half a second and then stops changing under you.

Not on this page, and on purpose: eighteen panels drawn three times over is a wall rather than a comparison, and the two questions a shortlist actually poses are answered better elsewhere. The domain appraisal puts up to three names beside each other with a figure on each, priced factor by factor and ranked on the estimate, with anything within a tenth of the leader called level rather than beaten. The brand preview puts the same two or three names side by side as a Google result, an app store listing, an inbox and a social profile, and marks no winner at all, because which of them looks like a company is the reader's call. Both carry the comparison in the address bar, so it is a link you can send, and both keep it while you come back here to look one of the names up in full.

From the domain itself, and every one of them is public. The certificate panel names the issuer, the expiry and every other hostname the certificate covers, which is public the moment it is issued and is often where a domain names its owner's other properties. The mail panel reads the domain's own DNS: the exchangers, the SPF policy, and the DMARC and MTA-STS policies, which live on names beside the domain rather than on it, at _dmarc and _mta-sts. DKIM is not there and cannot be: its keys live under a selector each sender chooses, so there is no name to ask about, and the panel says so rather than implying it is missing.

The answers are the same answers. What changes is that you get all of them for one query instead of retyping the domain into four boxes and holding four results in your head. Explore is the page for the open question, 'what is this name'; the individual tools are the pages for a question you already know you have.

From the domain's own sharing metadata — the og: and twitter: tags a site publishes so links to it render as a card, plus the page title, description and icon. That is the same data a messaging app or a search engine reads. Nothing is rendered or screenshotted, and no page content beyond those tags is stored or shown.

Because one set of tags renders five different ways, and a site missing one of them looks finished in some places and broken in others. Google prints the title and the meta description and ignores the share image entirely. X keeps the image and the domain and drops the words, unless the site asked for the small card, in which case it draws a thumbnail instead. Facebook stacks the domain in capitals above the title. LinkedIn discards the description rather than shortening it. iMessage puts the title under the picture. Each tab draws this domain's real metadata the way that platform would, so you can see which of them your link already looks right in — and no tab invents a field the site does not publish.

No. There are no affiliate links anywhere on this site, and we are not a registrar, a broker or a marketplace — the same footing as every third-party lookup in the actions row. The panel appears only for a name somebody already holds, and it names the route that fits the situation: a marketplace listing or a broker for a name in use, and a backorder for one whose term is nearly up or which the registry has already put in the deletion queue. It also states the two things those services do not lead with. A broker charges a fee whether or not the holder ever agrees to sell, and a backorder only pays off if the name is actually released — which usually does not happen, because the default at every registrar is to renew automatically.

It is not an audit. There is no crawl, no backlinks, no traffic estimate and no score out of a hundred, because those need data we do not have and a number pretending to summarise a site's prospects would be the least honest thing on the page. What it does is read the tags a search engine reads off the homepage — the title and description with their real lengths, the H1, the canonical URL, the robots directive, the language, the viewport, the structured data and the share card — and check each against the published limit, which is named in the tooltip. Those are facts rather than opinions, and four of them go wrong quietly: a noindex left on after a relaunch, a canonical still pointing at a staging host, a description written for a snippet length that has not applied for years, a missing viewport. The panel only appears when a page actually answered.

Because neither store has a handle namespace, so 'is this name claimed' is a question neither can be asked. Google Play is keyed on a package id nobody sees, and the App Store keeps whole titles unique rather than words. What they can answer is the question somebody naming a company actually has — how many apps are already called this — and every count here is worked out from the listings' own titles rather than taken from the store's result count, because both stores search by relevance and will happily return two dozen apps for a name none of them contains. A count with a plus sign means the search had more listings than were looked at, so it is a floor.

They are the names one mistake away from this one: a letter dropped, two letters swapped, a letter typed twice, an O read as a zero, a C spelled with a K, the plural, the hyphen closed up. Each is checked under this domain's own extension, because that is where a mistyped name lands. Which of them is registered is the useful part. On a name that is already somebody's brand, the taken ones are the neighbours catching its traffic. On a name you are about to buy, the free ones are the short list worth defending, and a lookalike already in somebody else's hands is a reason to think about the name again before you register it.

Not necessarily, and the panel is careful to say which it means. 'Nothing answered' is a statement about the check: the site may be down, may block automated requests, or may have no web server while still being used for email. 'No link preview' is different and is a real finding — a site answered and publishes no share card, which is why links to it post as bare URLs everywhere they are shared.

Because it could not, and saying so is the only honest option. Every panel depends on a different source, and any of them can rate-limit, time out or refuse an automated request. A panel in that state is reporting on the check, never on the domain, so it is never drawn as though the name or the handle were free.

No, and they are separate panels on purpose. The score reads the letters and answers three questions we can justify from the string alone: whether a first-time reader can pronounce it, whether somebody can type it again from memory, and whether it works as a brand at all given its length, its characters and its extension. It carries no price. The estimated value is the appraisal model, which needs things the string cannot give you — which extension the name is on, how much of the catalogue the word already occupies, how old the registration is — and answers with a figure, a range around it and how sure it is. Read the score for how the name will be spoken; read the value for what it would cost. Neither is an offer, and the full factor-by-factor working is on the appraisal page.

Extension availability is refreshed once a day, so treat it as accurate to within a day. The registration record, the DNS records and the handle checks are read live when you look. Whichever way, confirm with a registrar before you buy — we are not the registry.