Chinese, Japanese and Korean patents contain a wealth of chemical knowledge that can be unlocked by translating the chemical nomenclature. As an example, we applied our approach to 63k Korean patents (1990-2015) and extracted 1.7M distinct compounds. 231k of these compounds were not found by a comparable analysis of US patents. Of the 897k compounds, first disclosed between 2006 and 2014 by both a US and a Korean patent, for 415k the Korean patent was published earlier.