DeepL Voice now preserves your voice in real-time multilingual conversations
Language AI company DeepL has updated its Voice AI technology to preserve speaker tone and emotion in over 30 languages, alongside launching a new desktop application.
Language AI company DeepL announced major upgrades to its Voice AI technology on September 15, 2026, from Cologne, Germany. The company introduced upgraded Voice AI models and product enhancements engineered to deliver low-latency, real-time translation while maintaining the speaker's vocal identity. The updated voice preservation capabilities allow spoken translations to retain more of each individual speaker's unique tone, emotion, and natural expression during multilingual conversations.[1][2]
DeepL also debuted a dedicated desktop application designed to facilitate seamless multilingual online meetings. With this rollout, DeepL's voice-to-voice capabilities have reached general availability in more than 30 languages. The expansion builds on the platform's performance record, with DeepL Voice previously ranking first for translation accuracy in an independent assessment of AI-translated captions conducted by Slator.[1][2]
The company noted that DeepL Voice is scaling rapidly across enterprise deployments and major live experiences. Among its live implementations, the technology is powering real-time translation across all 50 stages at Salesforce Dreamforce 2026.[1][2]
