Unlock the past with Transkribus
Transkribus enables you to automatically recognise text easily, edit seamlessly, collaborate effortlessly, and even train your custom AI for digitizing and interpreting historical documents of any form.
100 credits for free. Every month.
Try it out
Tackle the Historical Document Dilemma
Navigate the complexities of historical documents with ease.
Transkribus is your AI-powered ally designed to simplify your time-consuming and laborious work with historical documents. With Transkribus, processes are automated, enabling you to effortlessly recognise text, and edit your materials. You can even train custom AI models for text and layout tailored to your specific needs. Additionally, Transkribus enhances your workflow by facilitating metadata enrichment, fostering collaboration, and giving you a powerful full text search. With Transkribus, you can streamline your work so you can focus on the fun stuff: unlocking the past.
150 000+
Registered Users
20 000+
HTR AI Models trained
50+ Million
Pages processed
150+
Free Public AI Models
Use AI-powered handwriting recognition
Using the power of AI, the Transkribus handwriting recognition software can be trained to decipher both handwritten historical documents and printed texts.
Find out more
Train your own custom AI models
Transkribus allows you to create custom AI models, specially trained to recognise the specific handwriting in your documents.
Find out more
Process large amounts of documents
Transkribus is designed for working with datasets of any size, small and large ones. Collections that were inaccessible before can now be read, searched, and prepared for further applications such as research, data analyses, or publishing.
Share your material with the world
Transkribus Sites makes it easy to share historical documents online and make entire collections searchable.
Find out more
Tag the text in your images
Tag words or phrases in your images, such as named entities, dates, events, but also places that are hard to read, etc.
More about Tagging
Leverage the search capabilities of Transkribus Sites
In addition to the full-text search option of Transkribus Sites, you can also enable Fuzzy Search (standard) or Smart Search (optional) to improve the search experience. Furthermore, the optional tag search lets your visitors search for specific tags.
See Transkribus Sites
Frequently Asked Questions
Transkribus is a comprehensive platform for the digitisation, AI-powered text recognition, transcription, and searching of historical documents. It has been developed during two EU-funded research projects and, since July 2019, has been maintained and further developed by READ-COOP SCE.
You can use Transkribus to: easily transcribe documents, both manually and automatically; train custom AI-models; search your transcriptions with powerful tools; annotate your documents with tags; work collaboratively; export the documents to various formats.
To get started, register on our website and access Transkribus here.
You can use Transkribus to: easily transcribe documents, both manually and automatically; train custom AI-models; search your transcriptions with powerful tools; annotate your documents with tags; work collaboratively; export the documents to various formats.
To get started, register on our website and access Transkribus here.
After having uploaded your documents, you can make use of public HTR models to automatically transcribe them. The public models have been trained by the Transkribus team and community and can be used by every user of the platform.
There is no general model for all scripts and languages, so you have to choose the best fit regarding language, script, and time period of your documents. On this page, you can search all the public models available and even test them directly.
If there is no public model that is suitable for your documents, you can train your own HTR model, as described here.
There is no general model for all scripts and languages, so you have to choose the best fit regarding language, script, and time period of your documents. On this page, you can search all the public models available and even test them directly.
If there is no public model that is suitable for your documents, you can train your own HTR model, as described here.
Whenever data are uploaded to Transkribus, they are stored on the servers of READ-COOP SCE (i.e., us, the company that develops and maintains the software). The servers are all located in Innsbruck, Austria, in a GDPR-compliant manner and the data may be processed by us according to the terms & conditions on our website. In short, this means that only personnel who work to maintain and improve the platform have access to the data, it will not be shared with anyone else and you retain full ownership and all rights to your data.
In case of personal data, you also have the right to have them deleted free of charge at any time. If you would like even more control and/or clarity, we invite you to conclude a data processing agreement with us. If you wish to do so, please send an email to info@readcoop.eu.
If you want to better protect your Transkribus account, you can add an extra layer of security with 2-factor authentication. After logging in, go to the Account page to enable this option.
In case of personal data, you also have the right to have them deleted free of charge at any time. If you would like even more control and/or clarity, we invite you to conclude a data processing agreement with us. If you wish to do so, please send an email to info@readcoop.eu.
If you want to better protect your Transkribus account, you can add an extra layer of security with 2-factor authentication. After logging in, go to the Account page to enable this option.
For more information on Transkribus, visit the Resource center on our website. Here you will find the manual, a glossary, many how-to guides, and videos to help you better understand all the features and become a successful Transkribus user.
Start unlocking the past with Transkribus
Leverage the power of Transkribus to get the most out of your historical documents.