DeepL wants real-time translation to finally capture sarcasm and mood shifts

DeepL is introducing a new translation model called Aure designed to capture nuances like sarcasm, emotional inflection, and speaking pace during real-time voice interactions. By preserving these delivery elements in the target language, the feature aims to move beyond literal word-for-word translation to convey the original speaker's intent more accurately. This update represents an effort to improve the quality of spoken communication in live, multilingual environments.