Skip to content
  • Pricing
jmgronski · PyLaia · Published January 7, 2026

Polski Wieszcz

Text Recognition

Description

Project description Polski Wieszcz is a general-purpose HTR model for historical Polish handwriting. Trained on documents from the 17th through 19th centuries, it is designed to perform well across a wide range of source types, including metrical records, notarial documents, and personal correspondence. Use it when the material is mixed, the document type is uncertain, or you need a strong starting point for training a more specialized model. Ideal for researchers, archivists, genealogists, and others working with diverse handwritten historical documents that include a broad Polish vocabulary. Credits Developed by the Polish team of the L'Dor V'Dor Foundation AI Lab, led by Jan Gronski, with generous support from JRI-Poland, the Polish State Archives, and individual volunteers Jacek Maryan and Jake Besser, who contributed documents and transcriptions. More information https://ldvdf.org/ai-lab/

Try this model

Drag an image here

Select a file...

PNG or JPG up to 10 Mb

Wolpi
AI Assistant

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

Polski Wieszcz
Use this modelOpen in Transkribus
Low error rate7.37% CER

Character Error Rate (CER) measures the percentage of characters incorrectly recognised. Lower is better. This model scored 7.37% on its validation set. As a rule of thumb, a CER below 10% is considered good for most handwritten material. This is a larger model trained on diverse material, which generally makes it more robust across different handwriting styles. That said, larger training sets also make it harder to push the CER down further.

Measured on the model's own validation data. Results on your documents may differ depending on handwriting style, document condition, language, and how closely your material resembles the training data.

Words298,276
Lines46,504
Training Pages1,280
Model ID464825
Languages
Polish