This addon predics all spoken text with the langident fasttext model ( https://fasttext.cc/docs/en/language-identification.html ). It's based on NVDAOneCoreAutoLang https://github.com/slohmaier/NVDAOneCoreAutoLang