This dataset provides detailed metadata on ca. 10.2 million works of fiction and non-fiction written after 1799 in 521 different languages. The dataset bolsters the May 2022 Hathifile by supplying missing fiction metadata with a bespoke BERT-based multilingual classifier.