How it works

Hearth doesn’t hold a database of houses. When you look up an address we query the public registers live (the land register, the heritage lists, the planning data platform, the energy register), match the records to your property, and assemble what comes back into one chronology.

That means the answer is as current as the registers are, and it means we can tell you what we checked even when the answer is “nothing”. A source with nothing to say about your house is still a fact about your house.

Land Registry writes your address one way, the energy register another, the council a third. Getting this wrong doesn’t produce a gap: it produces your neighbour’s history on your page, which is worse.

So every record is scored for how confidently it belongs to your building, and that score is shown. Confirmed means an authoritative register matched to your exact address. Possible means it might be the flat next door, and we say so on the record itself rather than burying it.

No free national source states when a house was built. The build-date estimate combines whatever signals exist (a new-build registration, the energy certificate’s age band, when the postcode was created) and tells you which ones it used.

When the signals are too weak or contradict each other, we say we don’t know. A range of “1700 to 1996” is not an estimate, and printing one would be a way of looking useful rather than being useful.

Census returns, probate calendars, newspaper archives and county record office catalogues have no API we can query. Nobody can fetch them for you programmatically: not us, not anyone.

What we can do is work out which archive holds the answer for your address and build the search for you, so you land on results instead of an empty search box. Some of those archives charge for access, and we earn a commission if you subscribe through our link. That’s disclosed every time, and it never changes which archives we recommend or the order they appear in.

There is no national, address-level, queryable dataset of planning applications in England. The national platform’s application dataset currently covers one authority and carries no addresses. Every council publishes its own register, across several different software platforms.

So we identify your council and send you to its register. That’s worse than a real integration and we’d rather say so than pretend. Where two energy assessments show your floor area changing, we flag that as probable building work and tell you roughly when, which is the one national signal for extensions that actually exists.

Hearth will not tell you who lives at an address now, and we don’t plan to build that.

Naming people at an address within roughly the last century means processing personal data about identifiable living people. It needs a lawful basis, it carries rights to object and to erasure, and the electoral register in particular has statutory restrictions on commercial reuse. Beyond the law, a searchable national index of who lives where is a stalking tool, whatever else it’s marketed as.

The historic record is different. Census returns are released after a hundred years, probate is a public calendar, and the people in them have been dead for generations. That’s where the interesting history is anyway.

The dividing line is drawn in code, not in policy: events flagged as naming potentially living people are removed before a response is built, whether or not the viewer has paid.

England and Wales. Scotland has a separate land register with different terms, and Northern Ireland a different archive system again. Both are on the list.