What Transkribus is
Transkribus is a web platform for turning historical documents into text. Its particular strength is handwritten text recognition (HTR) for Kurrent, Sütterlin, French cursive and other scripts that no conventional OCR engine can read. It is run by READ-COOP, a European cooperative. Processing happens on its servers in Austria. Users upload page images, choose or train a recognition model (there are hundreds of community models), correct the result in an editor, search across their collections and publish editions online. Models for printed text exist too.
It is paid in credits. On 15 September 2026 the pricing page listed 50 free credits a month, a Scholar plan at €99 a year for 900 credits and a Team plan at €449 a year for 1,500 credits, with one credit per page of handwritten recognition. For serious handwriting work a collection of a few thousand pages is a budget line rather than a flat fee.
What Searchive is
Searchive is a desktop application for printed text: scanned books, journals, typed reports, newspapers and registers that were printed and filled in with a typewriter. It recognizes German, French, English, Dutch and Italian on your own Mac or Windows machine, indexes every page and searches the whole library with the matches highlighted on the scan. There is no upload, no credits and no per-page cost. It costs a single payment and one document is free. It doesn't read handwriting and says so.
Side by side
| Feature | Searchive | Transkribus |
|---|---|---|
| Handwriting (HTR) | No | Yes, its core strength |
| Printed text | Yes, five languages | Yes, many models and languages |
| Fraktur | Not yet | Models available |
| Where processing happens | Your computer, offline | Transkribus servers in Austria (GDPR, European cooperative) |
| Correcting the transcription | No editor. Low-confidence pages are listed | Full editor, training your own models |
| Search | Whole library, offline, hit on the scan, three modes | Across your collections on the platform |
| Bulk import from a folder | Folders, subfolders, sets of photos | Upload per collection, batch tools exist |
| Cost model | Free for one document. Unlimited: €10 once at launch, €39 after | Credits per page, 50 free a month, plans from €99 a year |
| Data location afterwards | One folder on your disk | Your account on the platform, exportable |
| Collaboration, publishing | No | Yes |
Choose Transkribus when
- The documents are handwritten: letters, diaries, parish registers from before typewriters, notarial minutes, Kurrent and Sütterlin. No offline consumer tool does this well and Searchive doesn't do it at all.
- The print is Fraktur or another script that needs a specialized model.
- You want to train a model on a particular hand or typeface, correct transcriptions carefully and publish an edition.
- You work in a team on one collection.
Choose Searchive when
- The documents are printed or typed and there are thousands of pages: a shelf of journals, decades of reports, boxes of typed correspondence.
- The material may not be uploaded, whether because of an archive's terms, an employer's rules or your own judgment.
- You want the answer on the scan, fast, from one box, without credits ticking away.
- You work offline in the reading room, on the train or in a house without fiber.
Many researchers will need both. Handwritten sources go through Transkribus. The printed bulk of an archive, usually most of it by page count, goes into Searchive and is searchable the same afternoon on your own machine, free for the first document.
Join the waitlistOld print
Transkribus comes out of the historical-documents community, which is where its strength with old material comes from. Searchive was built by a historian and a programmer for one household's archive of twentieth-century journals and registers. It is tuned for printed text in modern typefaces, including old orthography, diacritics and worn print. Every page keeps a confidence score so you know which ones to check by eye. It can't read Fraktur or handwriting. Every page on this site says so instead of letting you find out after import. If you want Fraktur in Searchive, tell us through the waitlist. Demand decides what we build next.
Frequently asked questions
Can Searchive read handwriting at all?
No. Searchive recognizes printed and typed text. A handwritten page can be imported, viewed and pinned but its text isn't searchable. For handwriting you need handwritten text recognition, which is what Transkribus does.
Can I use Transkribus for printed documents too?
Yes, Transkribus offers models for printed text as well. The trade-offs are that pages are uploaded to its servers, processing uses credits and the result lives on the platform. Searchive does the printed part offline, in bulk, for a one-time price.
Does Searchive read Fraktur or Kurrent?
Neither. Fraktur is a printed typeface that needs its own recognition model, which Searchive doesn't ship yet. Kurrent and Sütterlin are handwriting. Transkribus has community models for both.
Is Transkribus a privacy risk?
Transkribus states that processing happens on servers in Austria under GDPR. It is run by a European cooperative, which is a much better position than an anonymous OCR website. It is still an upload. For some material an upload isn't allowed wherever the server is.