Twelve days to index 282,415 pages of 300-year-old French cursive.
Every parish register in a Swiss canton — all 1,097 volumes — read for one family, with every possible match opened and checked on the original page. It answered questions they had been stuck on for thirty years.
The records are public. That is not the same as findable.
Archives have been photographing their registers for years. A regional archive may have a thousand volumes online, free to anyone — and no way to search inside them. You can read any page you like, as long as you already know which page you want.
So the search narrows to wherever you already suspect an answer: the home parish, the likely decade, the spelling you expect. Everything outside that guess stays invisible. Families move. A clerk spells a name by ear. Somebody dies in a city nobody thought to check. And the record sits there for three hundred years, public and unread, because no one has the years it would take to turn every page.
That is the whole of it. Not a hard problem of interpretation — a brute problem of volume. And brute problems can now be handled.
Four steps. The third one is where the honesty is.
We read the whole archive, not the likely part
Not the home parish, not the expected years — every volume in the series, from the first page to the last. The point is to look where nobody thought to look, because that is where the missing people are.
We search for every way a clerk might have spelled it
Names were written down by ear, in a hurry, by people who had never seen them written. One family here appears as Remy, Remi, Rémy, Remie and Rémie. We search for all of it, and for the misreadings too, which is why we end up with thousands of possibilities rather than a tidy handful.
A person reads every possibility on the original page
Every one. On the scan of the actual register, in the actual handwriting. Most turn out to be somebody else, or nothing at all. That is the slow part, and it is not skippable.
What survives is written out with its source
The entry as the clerk wrote it, a translation into English, and exactly where it lives — which archive, which volume, which date. So you can quote it in a book, and anyone can go and check you.
Twelve days, start to finish.
The series runs to 1,097 bound volumes — 240 GB of page images, 282,415 pages. Read end to end at a reading-room desk, that is not a project. It is a career.
From the first volume to a finished, checked result was under twelve days. That is real elapsed time, false starts included — not a best case with the awkward parts left out.
Which changes what is worth attempting at all. Reading a whole archive used to mean a grant, a team and several years, so nobody ever did it for one family. Now the part that cannot be done by hand is the quick part, and the time goes where it should — into reading carefully and getting it right.
- Volumes1,097
- Page images240 GB
- Pages282,415
- Elapsed11 to 23 September
- Entries found86
- Registers cited38
Every volume is checked off before a series is called finished, so "we searched everything" means everything.
Eighty-six entries, in 38 different registers.
Most of them in parishes the family had never had a reason to open. A few of them:
You do not need to be technical. That is the point.
You have hit a wall
The paper trail stops, and the next record is somewhere in a series nobody has indexed. You are not short of skill — you are short of a way to look at a hundred thousand pages.
You are writing the book
A memoir, a family history, a village study. You need detail you can actually quote — what the clerk wrote, on what day — and a source note under every sentence.
You work with these records
A historian, a local society, an estate or an institution with a run of records and questions that would take a decade of page-turning to answer.
Rules we do not bend.
A person reads every page we cite
No fact reaches you because a computer thought it saw something. If it was not read by a human on the original, it does not go in.
Everything comes with its source
Archive, volume, date, and the words as the clerk wrote them — so you can quote it, and so anyone can check us.
We write down the dead ends
Where we looked and found nothing, with the exact years covered. It is the half of research that normally gets thrown away.
Something you can read, quote and hand on.
A written report. The findings set out properly, in plain English, with what is certain, what is likely, and what is still open — the thing you sit down and read.
Every entry, with its source. The original words, a translation, and the exact reference. Footnote-ready, whether you are writing a memoir or a journal article.
A private family website, if you want one. Your own photographs and papers indexed alongside the records, searchable, password-protected, and shareable with relatives who will never open a spreadsheet.
Everything we gathered. Yours to keep, so the next question does not mean starting over.
When we will tell you not to bother
The archive will not release the images. Some do, some do not, and some charge heavily. We check first, and if the pages cannot be got at, there is nothing to read.
The handwriting defeats it. Badly faded, heavily abbreviated or tightly bound pages come back as noise. Some hands and some alphabets are far harder than others.
The record was never kept. If the register burned, or the family changed its name on arrival, no amount of searching invents it. We would rather say so than take the work.
Stuck on a family that should be findable?
Tell us the name, the place, and where you have already looked. If there is a body of records that has never been indexed — a parish, a town, a regiment, a run of local newspapers — we will tell you honestly whether reading all of it would help, and roughly what that would take.
The family in this case study is our own, which is why we can publish the details. Client work stays private unless you say otherwise.