Within Transkribus, you can use many public models, already trained by the Transkribus community on a variety of scripts and languages
Previous step: Choosing a Model
Many AI text models are freely available in Transkribus. Some models have been created by the READ-COOP team, but most of them have been trained by Transkribus users, who have generously decided to share their models with the community.
On this page, you can browse an updated list of all the public models available in Transkribus and filter your search by century, language, material and script.
When you select a model, the page with its description and overview opens. Scrolling down, you can upload a test image and try the model out.
After having found and tested the best model for your document, you can use it in Transkribus to automatically transcribe your pages. In the "Text Recognition" window, just type its name or Model ID and select it.
If no model is satisfactory for your documents and you want to get better results, you can train your own AI text model, as explained on the Training Text Recognition Models page.
Next step: Super Models