The QVIRE project is part of the joint initiative of Dr. Evina Stein, Comenius University in Bratislava, and the Kempelen Institute for Intelligent Technologies. Our goal is to lay the foundations for the development of Cultural AI in our region and to help establish the first Cultural AI lab in Slovakia. This multidisciplinary project aims to bring the expertise in artificial intelligence (AI) to the cultural heritage sector and humanities research to Slovakia.

Currently, no institution or organization in Slovakia actively engages in the development of AI solutions for written cultural heritage. For Slovakia, this is a promising opportunity to become a major player in this emerging field of applied research in the future and an opportunity to join the European leaders in this area.

Dr. Evina Stein is an expert on medieval Western manuscripts currently working as a guest researcher at the Huygens Institute in Amsterdam. Dr. Stein is a promising Slovak early career researcher with an ambition to obtain an ERC Starting grant to start a research group in Slovakia. Comenius University in Bratislava and KInIT have teamed up to assist Dr. Stein in this endeavor. 

In addition to the resources provided by the QVIRE project, Dr. Stein has also received support from the mentoring program of the ESET foundation aimed at supporting ERC applications.


Medieval manuscript books are an important component of European cultural heritage. They serve as valuable sources of data for many types of historical research. One of the greatest challenges to using manuscripts in historical research is that we typically do not know when, where, and in what historical context they were produced, nor where they were used. 

This is the reason why, for more than three centuries, paleographers and codicologists, two categories of manuscript specialists, have been developing methods to quickly, efficiently, and accurately date, locate, and contextualize manuscripts based on their intrinsic features (e.g., script and material). Their expertise has become invaluable for our ability to critically evaluate the historical manuscript data. 

The major obstacles to traditional paleographic and codicological assessment of manuscripts are:

  • the importance of large datasets for reaching significant conclusions (as small case studies tend to provide results of limited significance)
  • its time-intensive, expertise-intensive, and costly character, as the assessment is typically based on in situ examination by highly-trained experts 
  • potentially high level of human error or subjective perception

Many of these obstacles can now be partially removed with the help of artificial intelligence, in particular as many medieval manuscripts are now digitized online.

The teaming up of AI and traditional codicology has a great potential to significantly advance historical research on European cultural heritage. Specifically, AI may make it possible for manuscript researchers to harvest larger manuscript-based datasets from the digital manuscript facsimiles than has been previously possible, and to do so with greater precision and at a fraction of time and cost it would take a human research team. 

The QVIRE project aims to explore one avenue for the potential of AI-powered historical research by creating an artificial intelligence model and a software prototype for efficient automated data collection and analysis of one feature of historical manuscripts, the quire marks. Quire marks were used to guide the medieval manuscript manufacturers physically assembling a codex. They are present in almost all medieval manuscripts, but have rarely been used for their dating, localization, and contextualization due to the difficulties posed by their big data analysis. While it is unlikely they could be ever analyzed systematically by traditional means, they are an excellent subject of examination with an AI model.

A necessary basis for the creation of such a model is the selection of an appropriate digitized manuscript corpus and an expert annotation of a large amount of digital manuscript images. The data selection and annotation will be carried out by project investigators with expertise in working with historical manuscripts. The tools developed within this project will be accessible to a wide professional public, especially professionals from the heritage sector (i.e. staff of museums, galleries, archives, libraries, etc.).

Main partners and collaborators

Dr. Evina Stein

Dr. Evina Stein is a guest researcher at the Huygens Institute in Amsterdam and an expert on medieval manuscripts.

Prof. Juraj Šedivý

Prof. Juraj Šedivý is Head of the Department of Archive Studies and Museology, Faculty of Philosophy, Comenius University and an expert on medieval documents. He is the only European historian who is a member of three global commissions in historical sciences (International Palaeographical Committee, International Commission on Diplomacy and International Commission on Urban History). 

Martin Tamajka

Martin Tamajka is a researcher and lead engineer for artificial intelligence at the Kempelen Institute for Intelligent Technologies, he is an expert on image processing.  

Marián Šimko

Marián Šimko is Head of a research team focused on natural language processing, which includes multimodal content processing, at Kempelen Institute for Intelligent Technologies.


  1. A publicly available artificial intelligence model for the automatic detection of quire marks in medieval manuscripts, which enables the analysis of large corpora for the purpose of historical research (model, data).
  2. A publicly accessible software tool prototype for the analysis of quire marks, based on an artificial intelligence model developed within the QVIRE project (online tool).
  3. A publicly accessible scientific webinar. The webinar will present the model and the software prototype developed within the QVIRE project. We will explain the various options for using artificial intelligence in the humanities and their potential for making research in this field more effective.
  4. Presentation at a scientific conference – Publication of project results at the international conference Novel approaches to Digital Codicology. We introduced the procedures, findings and conclusions obtained during the project to the professional audience.
  5. This project website.


  • Webinar “Project QVIRE: AI aids medieval manuscript research” (29 March 2023)

This project was realized with the financial support of the Special Fund of the Prime Minister of the Slovak Republic.