How this registry is built, and what it refuses to do

Most sites about cars are written. This one is derived. The difference is not a stylistic preference — it is the reason a figure on this site can be checked, and the reason there are far fewer pages here than a content operation would produce.

A page is not created because someone wrote it

Every model-in-market page has to earn its own address. A scoring gate examines what is actually known about that pair — confirmed availability with a source, a local price list with dates, used market data, tax treatment, dealer coverage, signed local editorial, local press tests, differing trim names — and only a record that clears the threshold becomes a page. A weaker record becomes a tile on a brand hub. A record below that gets no presence at all.

This runs on every build. Nobody can override it by writing more text, because text is not what it counts.

The rule exists because publishing thousands of near-identical pages, each differing only in a country name, is the pattern search engines penalise, and the only honest defence is a mechanism that refuses to publish when there is nothing locally different to say.

No figure is ever typed into a sentence

Prices, power, range, dealer counts — none of them appear in editorial prose anywhere on this site. They live in data fields and are rendered by components. A separate check scans every editorial paragraph on every build and stops the build if a number leaked into one.

The reason is repair. A wrong figure in a data field is corrected once and every page that shows it is correct from that moment. A wrong figure written into a sentence spreads through the site and cannot be found again.

Every local field carries a source and a date

A price without a link to the importer’s published list does not exist here. Neither does an availability claim, a tax figure or a dealer count. The absence of a source is not a gap to be filled with an estimate — it means the field is empty, the page scores lower, and it may not be published at all.

Where an importer has published no price list, the page says so in plain words. It does not estimate, convert from another market, or infer from a neighbouring country.

Prices are cross-checked against a second source

An importer’s published list says what a car costs on paper. A cross-check against the manufacturer’s configurator or a dealer quote says somebody has seen that figure in the real world. Where a cross-check exists, the page names it and dates it.

Where the two disagree, the page says that too. A gap between a list price and a configurator is information the buyer needs, not an embarrassment to hide.

An expired price list does not delete the page

When a list passes its validity date, the page keeps its figures and shows the expiry date prominently instead. A page that vanishes from search results because research fell behind is worse for a reader than a quarter-old price carried with an honest note.

Who runs this register

Research here is carried out by an AI system working against manufacturer and importer material. There is no staff of human editors, and this site does not pretend otherwise: what stands under a verdict is an editorial signature, not an invented name. A person who does not exist is not a weaker trust signal than a real one — it is a false one, and a register whose whole argument is verifiability cannot afford to open with a lie about its own authorship.

It is worth naming what that trade actually costs, in both directions. The system reads a price list, a brochure and a type-approval document faster and more consistently than a person, and it never forgets to record the URL it came from and the date it was taken — which is why every field on this site has both. What it cannot do is walk into a showroom. It will not hear what a salesperson says, will not notice that the printed price list on the counter differs from the one on the importer’s website, and will not know that a version quietly stopped being orderable last month.

Only a reader who was there knows that. So every model page carries a correction form, and what comes back is treated as data rather than as feedback: an accepted correction is written into the field with its date, exactly like any other source. This is the one input the register cannot gather on its own, which makes it the one place where a reader changes what the site says.

Tax and customs sections are never translated. They are written from scratch for each jurisdiction with a local source, or they are absent. And they describe a process — never any particular person’s liability.

Language and country are different things

A glossary entry or a brand article is written once per language and serves every market that speaks it. It is not duplicated per country, because duplicating it would create twin pages inside one language — the exact pattern the whole architecture exists to avoid. A separate check compares markets sharing a language on every build and stops the build if their pages have grown too similar.

What belongs to a country is price, availability, tax treatment, dealer coverage and used supply. Nothing else.

Pages enter the index in cohorts

Every page is built as soon as its data is ready, but it enters the search index only when its publication wave is released. A fresh domain that drops hundreds of addresses in one day looks like what it is. Releasing in cohorts of roughly thirty addresses per market per month does not.

A model outside a released wave carries a noindex directive and disappears from the sitemap, from language declarations, from the finder, from categories and from comparisons — because a comparison linking to a page excluded from the index would be a dead end.

What we do not do

We do not sell cars and we are not a broker. We do not create a page per city or a page per search phrase variant. We do not estimate a price the importer has not published. We do not open an enquiry form for a market where nobody is on the other end. We do not carry a Chinese-market range figure and hope the reader notices the difference in test cycle.

Corrections

When a figure here turns out to be wrong, it is corrected in the data field and the change reaches every page that shows it. The form sits at the bottom of every model page. Tell us what looks wrong and where you saw the correct figure — a source with a date is worth more to us than a report without one, because a source can be checked and a report has to be taken on faith.

Nothing changes on the strength of a submission alone. Every correction is verified against the source before the field moves, which is the same standard the original research is held to.