• I translate, say, English to Danish on deepL;
• all the translations are put on a temporary vim buffer, then do be pasted on Sound Of Text); this website generates an audio file with pronunciation;
• for each term to be translated, I find the source (usually in English), target (the translation) and pronunciation audio file — the translation trifecta and I open the Anki card creation panel, with a file-browser on the side. Then, the Anki flashcard form is filled manually.

This, however, takes a lot of time.