Keymakr contributes to RUKOPYS, the first open dataset of Ukrainian handwritten text
Summary
Keymakr became the official labeling partner for RUKOPYS, the first comprehensive open dataset of Ukrainian handwritten text covering over 100 years of documents. Keymakr's in-house experts performed line-level annotation, transcription, normalization, and multi-stage quality control to convert varied, degraded, and historical handwriting into training-ready data for OCR and HTR models. The dataset, anonymized and released under an open license, supports the Handwritten to Data AI Challenge and offers AWS access for participants, with potential integration into Ukraine's ePermit and archival digitization workflows. The project advances Ukrainian-language AI resources, preserves cultural heritage, and demonstrates Keymakr's capabilities in building high-quality training data pipelines for public-sector and research applications.