You have the sources but can't search them.
You have folders of scanned journals, church registers, correspondence and city records: thousands of pages you photographed in archives or inherited as PDFs. Finding one name in them means reading all of them. Cloud OCR services could do it, but they want your documents uploaded to someone else's computer.
Import. Recognize. Search.
Import PDFs or photos
Single files, whole folders or the photos you took in the archive: a set of phone photos becomes one document with the pages in camera order. Your originals are never modified.
It reads the scans
Text recognition set per document, in the language the document is written in, including old orthography and diacritics.
Find any word, on the page
Full-text search across every page in your library, with matches highlighted on the scan itself so you read the original and not a transcription. Search all words, any word (Cöln, Köln and Coeln in one go) or an exact phrase.
A whole library behind the search box.
Pin the pages that matter with a note in your own words and collect your findings in one place. Export any search's results as a CSV for your notes or source list. Projects group your sources, per-page read-quality flags show which pages to double-check and metadata always says whether the document stated something or the app inferred it.
Your archive never leaves your machine.
The documents, the recognized text and the search index all live in a folder on your own computer. Nothing is uploaded, no account is needed and there are no analytics. No model reads or learns from your sources. It works with the network cable unplugged.
No AI reads your archive
Recognition is a fixed model that ships inside the app. Nothing is uploaded, so nothing can train on your documents or end up in anyone's model. How to check that yourself.
Works fully offline
In the archive basement, on the train, anywhere. Nothing about it needs a connection.
Your files stay files
One ordinary folder holds everything. You can back it up, move it and open it without a proprietary vault or any lock-in.
Yours, permanently
Pay once and it is yours. There is no subscription to lapse and no renewal to remember. Your library keeps working.
Start free. Pay once when you need more.
Free
- One document, as many pages as it has
- Every feature, nothing held back
- No account or card needed and no time limit
Searchive
€39 after launch.
- Unlimited documents
- Every feature, all five languages
- Pay once, use it for as long as you like
Waitlist members get the €10 launch price by e-mail when we open. No card now. After that the price is €39, paid once. The free plan stays free: the full product on one document, for as long as you like.
Built by the people who needed it.
Searchive began as one Dutch household's problem: a historian with thousands of scanned pages (journals, registers and correspondence in four languages) and no way to search any of it. The cloud services that could read them wanted the archive uploaded first. A century of papers on someone else's servers felt wrong. So the programmer across the table built the tool they couldn't find, one that is fast, complete and entirely on your own machine.
That's still the whole company: a researcher and an IT engineer in the Netherlands. There is no support department. Mail sent to info@searchive.io is read and answered by us.
Questions
Is this a subscription?
No. You pay once and keep using it, with no monthly fee and no renewal. Before you buy you can try everything for 14 days. After that the free plan keeps one document fully working and the rest of your library stays readable, searchable and exportable.
Does it work offline?
Yes. Recognition and search run entirely on your machine. A paid license is confirmed with a brief online check that sends your license key and nothing else, so no document data, filenames or search terms. The app then works offline for weeks at a time.
There are free tools that index files. Why pay?
Free desktop indexers tell you a word occurs somewhere in a file and leave you to open it and hunt for the page. Searchive opens the scanned page itself with every match highlighted on the image. Historical text is also full of the characters those tools stumble over: names like l'église, Sa'ud or d'Annunzio, accents and quotation marks. Searchive finds them typed with or without the apostrophe or the accent. Tests cover that behavior. Photo import, projects, pinned pages with your own notes and CSV export of your findings make it a research tool as well as an index.
I only have photos, not PDFs. Does that work?
Yes. Select the photos of a document and they become one searchable document, pages in the order the camera took them. JPEG, PNG, TIFF and WebP are supported.
Which languages can it read?
German, French, English, Dutch and Italian, chosen per document. A French letter and a German newspaper in the same library each get read in their own language. More languages are planned.
How good is the recognition on old print?
Good enough to search by: clean twentieth-century print recognizes nearly perfectly and every page gets a confidence score so you can see which pages deserve a second look. Handwriting is not supported.
What does it run on?
macOS (Apple Silicon) and Windows 10 or later, 64-bit. It is a normal desktop application with one installer and no dependencies to set up.
Where exactly is my data?
In a single folder you choose, holding your originals, the searchable copies and the index database. Point your backup tool at it and everything is covered.