DeepL has launched Aure, its most dynamic and realistic voice AI translation model to date—though Chinese is not yet supported. As the underlying engine powering DeepL Voice, Aure can dynamically adapt translated speech, reproducing in real time the speaker's tone, emotion, stress, and pace, fully preserving the naturalness and expressiveness of human conversation. Building on DeepL's existing voice preservation technology, the model retains the speaker's timbre while further capturing the dynamic vocal fluctuations and split-second reactions that occur in conversation, imbuing real-time translation with the spontaneity and authentic spirit of natural human speech. Aure currently powers DeepL Voice for web meetings, face-to-face conversations, and the DeepL Voice API, initially supporting English, German, Spanish, French, Italian, Japanese, Portuguese, and Korean, with more languages to be added progressively. Languages not yet supported will continue to use DeepL's previous voice model.