Transkribus is a platform for automated transcription of historical manuscripts and prints. Developed at the University of Innsbruck, it offers an array of tools and support provided by its community, which is currently made up by over 100 000 users, continually releasing new models and making them available for subsequent use.

Transkribus at the OeAW

Transkribus is already successfully being used at several institutes of the Austrian Academy of Sciences for text- and layout recognition. Researchers from numerous disciplines working in different project contexts are using various components of Transkribus' tool suite. The ACDH connects Transkribus users at the OeAW: Basis for this network is the knowledge transfer between OeAW institutes, an exchange about the most recent developments of the platform as well as interdisciplinary reflection of the use of AI in the field of Digital Humanities.

 

Transkribus Coordinators at the OeAW

Robert Klugseder

Austrian Centre for Digital Humanities
Department of Musicology

Claudia Resch

Austrian Centre for Digital Humanities
Department of Literary & Textual Studies

News

At the beginning of the year 2024, the Read Coop has offered the Austrian Centre for Digital Humanities 30.000 free Transkribus Credits. The ACDH particularly encouraged young scholars to propose transcription projects to be supported with these credits. Applicants should be willing to work according to the FAIR principles in order to qualify for the credits.

The closing date for submissions was 15 March 2024. In April the credits have been awarded to:

  • William Godsey / IHB: 5.000 credits for transcriptions within the project "Dynastic Continuity in an Age of Crisis? The Habsburgs in the Revolutionary and Napoleonic Wars (1792-1815)"
  • Jan Odstrcilik / IMAFO: 5.000 credits for transcriptions of the sermons of Heinrich Totting von Oyta
  • Yasir Yilmaz / IHB: 1.500 credits for transcriptions within the project "The Ottoman Grand Vizierate 1560s to 1760s (GraViz)"

 

Transkribus Events

Upcoming: 

There are currently no upcoming events.

Archive: 

 

Transkribus for students

Apply for a Transkribus CLARIAH-AT Scholarship.

 

Transkribus Projects at the OeAW

The Accounts of Aldersbach Monastery
Team: Robert Klugseder (PI), Magdalena Rufin, Johanna Maria Unterholzner
Institute/Institutions: ACDH, Universität Graz, Know Center Graz, Universität Regensburg
Financing: FWF, DFG (Weave program)
Project duration: May 2024–Dec 2027

Digital edition and analysis of around 20 German and Latin account books from the 15th and 16th centuries. Automatic text transcription with Transkribus with layout and text model training.

 

ARITHMETIC - German Arithmetical Treatises in Manuscripts of the Late Middle Ages (1400-1522)
Team: Michaela Wiesinger (PI), Christina Jackel, Norbert Orban, Franziska Putz, Bernhard Bauer
Institute: IMAFO
Financing: ERC
Project duration: Until September 2027

We are working on an "assertive edition" of the texts. Our roughly 140 primary sources are transcribed using Transkribus where we generate our own HTR-models. The first annotations are also done with this software before we export our xml-files for further processing.

 

Arthur Schnitzler - Critical Edition (Early Works) IV
PI: Konstanze Fliedl
Institute: ACDH
Financing: FWF
Project duration: Until May 2024 

For the historical-critical edition, primarily relevant printed texts (first prints and first editions) as well as materials from the estate are read in. While prints and typescripts are OCR read and then collated, manuscripts are manually transcribed in Transkribus. The texts are then exported as xml.

 

Arthur Schnitzler: Collection of newspaper clippings
Team: Peter Andorfer, Martin Anton Müller, Ingo Börner
Institute: ACDH
Financing: various
Project duration: Sept–Oct 2019

Automatic recognition of microfiche scans. The website currently mirrors the Transkribus server, including the images, via the API.

 

Auden in Austria Digital (AAD)
Team: Sandra Mayer (PI), Timo Frühwirth
Institute: ACDH
Financing: FWF
Project duration: 2024–2027

Training of a model (to be made public in the future) for the 'later' (i.e. in the period 1958-1973) handwriting of W. H. Auden and application of public models for typewriting and printing in English and German.

 

Clara Katharina Pollaczek: Arthur Schnitzler und ich.
Team: Martin Anton Müller (PI), Laura Untner
Institute: ACDH
Financing: FWF
Project duration: Nov 2022–Jan 2023

Automatic recognition of the typescript. Crowd sourcing for corrections in Transkribus. Automatic workflow for export to a static page.

 

Classical Text Editor
Responsible: Stefan Hagel
Institute: ÖAI
Financing: self-financing
Project duration: unlimited

The Classical Text Editor facilitates producing critical editions both in printed and digital form; it imports texts from Transkribus along with the links to image regions.

 

DWA-Austria – pilot study
Team: Alexandra N. Lenz, Markus Kunzmann
Institute: ACDH
Financing: Pilot study via OeAW, preparations of FWF proposal
Laufzeit: Oct 2022 ­– Sept 2023

Full text digitization of Austrian survey forms for the German Word Atlas/Deutscher Wortatlas (DWA).

 

Edition of Arthur Schnitzler's literary correspondences
Team: Martin Anton Müller (PI), Laura Untner
Institute: ACDH
Financing: FWF
Project duration: 2018–2024

Transcription of pieces of correspondence. Training of free manuscript models.

 

Eduard Hanslick’s Criticism between Aesthetics, Journalism, and Scholarship
Team: Alexander Wilfing (PI), Katharina Bamer, Daniel Elsner, Christoph Landerer, Anna-Maria Pfiel
Institute: ACDH
Financing: FWF
Project duration: May 2022–Oct 2025

In the Hanslick Online project, Transkribus is/was used for automated text recognition and the data is then converted into a TEI/XML format via API.

 

Family Matters. Female Dynastic Agency and Correspondence
Team: Katrin Keller (PI), Peter Andorfer
Institutes: IHB, ACDH
Financing: FWF
Project duration: 2021–2024

In this project, Transkribus was primarily used to facilitate transcription. We realized the data import and export as well as the data conversion to TEI/XML using the Python package acdh-transkribus-utils developed at ACDH-CH.

 

Listen in historischen Zeitungen: Herausforderungen und Potenziale der digitalen Analyse einer vernachlässtigten Textsorte
Team: Nina C. Rastinger
Institute: ACDH
Financing: OeAW DOC grant
Project duration: 2023–2026

The dissertation project, which is funded by an OeAW DOC grant, uses methods of digital text science and corpus linguistics to analyze periodically published lists in printed German-language newspapers between 1600 and 1850. Transkribus will be used to create full-text digital copies of selected lists and to test strategies for systematically identifying the type of text in historical newspapers, e.g. on the basis of their layout.

 

Making the Austrian Federal Constitution 1920
Team: Thomas Olechowski, Jörg Kammerhofer, Alexandra N. Lenz
Institution/Institute: Universität Wien, ACDH, Universität Freiburg
Financing: FWF, DFG
Project duration: Jan 2023–Dec 2025

There are relatively many textual witnesses - partly drafts of the constitution, partly "paratexts" such as minutes of parliamentary sessions etc. - which are quite diverse in terms of material; e.g. typescripts, manuscripts, prints and lithographs. The scans are uploaded from Goobi to Transkribus for basic text recognition and corrections. To capture/encode typographical, spatial and genetic information only XML, not the features offered by Transkribus, is used.

 

Oberkammeramtsrechnungen (accounts) of the City of Vienna
Team: Robert Klugseder (PI), Daniel Schopper, Fernando Sanz-Lázaro, Johanna Maria Unterholzner and Peter Andorfer
Institutions: ACDH, Wiener Stadt- und Landesarchiv (MA 8)
Financing: BMKÖS (Kulturerbe Digital)
Project duration: Nov 2024–April 2026

Digital edition of 289 German account books from 1424–1768. Automatic text transcription with Transkribus with layout and text model training. The accounts records contain important information about the history of Vienna.

 

Plainchant manuscripts of the Central Library of the Viennese Franciscan Province Graz
Team: Robert Klugseder (PI), Daniel Schopper
Institute: ACDH
Financing: BMKOES
Project duration: June 2023–August 2024

The aim of this project is the digitisation of 53 late medieval and early modern chant manuscripts, chant organ books and rare chant prints from the Central Library of the Franciscans in Graz. Automated transcription of the lyrics, training and application of models for layout and text writing.

 

Recognition of musical notation in medieval and early modern chant books
Team: Robert Klugseder (PI), Jorge Calvo-Zaragoza, Schorsch Hackl
Institute: ACDH
Financing: ACDH, Universidad de Alicante, READ COOP SCE
Project duration: Dec 2023–Aug 2024

Automated transcription of song text and music notation in liturgical chant manuscripts. Development of an extension for Transkribus, which will make it possible to automatically transcribe medieval plainchant.

 

TibSchol – The dawn of Tibetan Buddhist Scholasticism (11th-13th c.)
PI : Pascale Hugon
Institute : IKGA
Financing : ERC
Project duration : Until June 2026

A layout analysis model (“Tibetan Pecha”) and two HTR models have been trained on the Transkribus platform for the transcription of documents in two varieties of handwritten Tibetan cursive script. These models have been made public in Transkribus and the Ground Truth is available on GitHub. Links to the documentation, datasets and pre-processing tools can be found in the “Resources” section of the project’s website (https://www.oeaw.ac.at/projects/tibschol/publications-and-resources).

 

Vienna's memorable sites – Deep mapping early modern travel guides
Team: Nina C. Rastinger (PI), Stefanie Salzburger
Institute: ACDH
Financing: Stadt Wien
Project duration: 2023–2024

The project, funded by the City of Vienna, analyzes historical perceptions of Vienna by combining travel guides from the 18th and early 19th centuries with digital text analysis and GIS-based methods. Transkribus is used to generate reliable full texts and to carry out (semi-)automatic structural annotations, the latter in particular by training specific layout recognition models for each work.

 

Visiting Vienna – Digital approaches to the (semi-) automatic analysis of the arrival lists found in the "Wien[n]erisches Diarium"
Team: Nina C. Rastinger (PI), Thomas Kirchmair
Institute: ACDH
Financing: Stadt Wien
Project duration: 2022–2023

As part of the project funded by the City of Vienna, Transkribus was used to create reliable full texts of the arrival lists printed in the Wien[n]erisches Diarium, which were then semantically enriched using Named Entity Recognition methods and mapped onto historical city maps of Vienna.

 

Wien[n]erische Diarium digital
Team: Claudia Resch (PI), Dario Kampkaspar, Nina C. Rastinger
Institute: ACDH
Financing: ÖAW go!digital, Stadt Wien, ACDH
Project duration: 2017 ff.

In this project, Transkribus was applied for the first time to a large dataset of Fraktur printing for the full-text indexing of the historical Viennese newspaper. One result of this is the public AI model "Deutsche Fraktur 18. Jahrhundert", which is now used in many digitization projects.