DeepL says Aure goes beyond reproducing a fixed voice sample from the start of a conversation to reflect changing tone, emphasis, emotion and speaking pace. Initial support covers eight languages including Korean, with automatic application to online meetings, in-person conversations and the API. DeepL aims to expand into large events using low-latency transmission expertise from the Mixhalo team. This is not independent proof that emotion or sarcasm is always translated correctly, and maintaining consistent latency at large events remains an ongoing challenge.