You might be able to build something similar yourself using declension data extracted from Wiktionary using wiktextract: https://github.com/tatuylonen/wiktextract#pre-extracted-data