The Future of Information Extraction – Be Part of TUC 2024! ✨ Feb 15-16, In-Person and Online. Get your Ticket >>

Ukrainian generic handwriting

Free Public AI Model for Handwritten Text Recognition with Transkribus

Ukrainian generic handwriting

This first version of a generic model for handwritten Ukrainian (predominantly from the second part of the 19th to the late 20th century) was curated and trained by Aleksej Tikhonov, a researcher of the MultiHTR project (Freiburg/Germany, PI: Achim Rabus).

Portions of the GT data have kindly been provided by Kateryna Lobuzina of The Vernadskyi National Library of Ukraine (VNLU), containing manuscripts by Taras Shevchenko, as well as Klyment Kvitka’s notices and comments on Ukrainian folklore collected by Lesya Ukrainka. Another part of the GT has kindly been provided by the corpus created by Maria Shvedova, Ruprecht von Waldenfels, Serhij Yaryhin, Andriy Rysin, Vasyl Starko, Tymofij Nikolajenko, et al. (2017-2023): GRAC: General Regionally Annotated Corpus of Ukrainian, Electronic resource: Kyiv, Lviv, Jena (, which contains a collection of letters prepared by students of Lviv Polytechnic University in 2018-2020 (

We thank Mykhailo Kostiv of The National Museum of the Holodomor-Genocide in Kyiv, who provided the training process with a manuscript by Lavrin Nechyporenko from the 1960s.
We sincerely thank our partners from Ukraine for the cooperation with MultiHTR on the first generic model of handwritten Ukrainian, despite the ongoing war in their country.

Image source: Vernadskyi National Library of Ukraine.

Model Overview

Ukrainian generic handwriting 1
MultiHTR project (Achim Rabus & Aleksej Tikhonov)
Model ID:
19th, 20th
Cyrillic alphabet
CER on validation set:
4.20 %
Simply upload a picture and test this model

By uploading an image, you accept our terms and conditions and our privacy policy

Ukrainian generic handwriting 1 is freely available to everyone

Get started with Transkribus and use it for your own Material
You can use this model to automatically transcribe Handwritten documents with Handwritten Text Recgnition in Transkribus. This model can be used in the Transkribus Expert Client as well as in Transkribus Lite.
This AI model was trained to automatically convert text from images of historical Cyrillic alphabet documents into editable and searchable text.