DeepL launches real-time voice translation "DeepL Voice" to break down language barriers

DeepL Japan held a press briefing for DeepL Voice, a service designed to support smooth, barrier-free communication in both in-person and online settings.

DeepL Voice has two offerings: DeepL Voice for Meetings, a browser-based AI captioning solution that integrates with Microsoft Teams and Zoom and provides real-time translation and transcription; and DeepL Voice for Conversations, a smartphone-based real-time subtitle solution available via mobile app and web browser with conversation and split face-to-face display modes and customization options.

Aimed at business use, the product emphasizes high translation accuracy and readability compared with typical web-conference translators, handles Japanese sentence complexity, reduces fatal mistranslations and subtitle flicker, suppresses hallucinations, and includes immediate deletion of translation data (not used for AI training), encryption, and GDPR compliance.

Corporate trials reported productivity gains: about 194 hours saved per person per year and roughly ¥960,000 in annual cost savings; meeting preparation and follow-up work fell by 26%, and real-time transcription streamlined minutes creation.

Voice-to-Voice translation (preserving speaker voice) is under development with a 2026 target including Japanese support; DeepL API was presented (one input language can be translated into up to five target languages simultaneously), and a Notion case showed reduced language barriers, seamless in-app translation, and lower operating costs.