DeepL, the translation tool that’s been quietly eating Google Translate’s lunch for years, just announced its move into a whole new arena: real-time voice translation. The company unveiled DeepL Voice-to-Voice, a suite of services designed to translate spoken conversations on the fly.
DeepL Voice – A secure voice translation tool for global teams
OpenAI unleashes GPT-5.4-Cyber, a high-clearance AI for cybersecurity pros
Alibaba's new Qwen3.6 model wants to be your AI coding partner
Anthropic's Claude Opus 4.7 arrives with smarter coding, better vision, and tighter safety ）
OpenAI unleashes GPT-5.4-Cyber, a high-clearance AI for cybersecurity pros
Alibaba's new Qwen3.6 model wants to be your AI coding partner
Anthropic's Claude Opus 4.7 arrives with smarter coding, better vision, and tighter safety
DeepL unveils real-time spoken translation, breaking the next language barrier with Voice-to-Voice
**DeepL, known for text translation, now wants to translate your voice | TechCrunch
On April 16, 2026, DeepL announced DeepL Voice-to-Voice, a set of features built specifically for real-time spoken communication. By expanding into voice translation, DeepL aims to provide instant, high-accuracy translation via API for video calls, in-person conversations, customer support, and more—helping teams collaborate without worrying about language barriers.
Here’s what’s included in the DeepL Voice-to-Voice suite.
DeepL Voice for Meetings
This feature brings real-time translation to platforms like Microsoft Teams and Zoom, letting users speak in their native language while other participants hear it in theirs. DeepL Voice for Meetings is slated to start an early access program in June 2026, and sign-ups are open now.
DeepL Voice-to-Voice Early Access Programme
DeepL Voice for Conversations (mobile and web)
DeepL Voice for Conversations expands beyond mobile to offer a truly multi-platform experience, deployable even in environments where installing an app isn’t practical or permitted.
Group Conversations
A feature for settings like training, coaching, or workshops, where participants can join instantly via a QR code to enable multilingual discussion. This lets on-site employees maintain a shared understanding during hands-on interactions with multiple speakers. Because it’s accessible from multiple devices, participants receive simultaneous, real-time voice translation. The group conversation feature will be available to the public starting April 30, 2026.
Voice-to-Voice API
Lets businesses integrate DeepL’s voice translation directly into internal applications or customer-facing tools like contact centers. An early access program is currently open for registration.
Customization with Voice Terminology
A new quality optimization feature for DeepL Voice that ensures industry-specific terms—product names, company names, people’s names, you name it—are captured, transcribed, and translated more accurately in real time, even if the speech is fast or packed with jargon. As part of this, DeepL’s translation glossary integrates with DeepL Voice, allowing users to standardize key terminology across an entire conversation. It launches publicly on May 7, 2026.
Most existing voice translation systems follow a clunky three-step process: speech to text, translate the text, then text back to speech. DeepL argues its years of experience in text translation give it a leg up on quality. But the company says it’s aiming even higher: eventually developing an end-to-end voice translation model that skips the text conversion step entirely.
TechCrunch notes a potential direct competitor: Palabra, a real-time voice translation engine designed to preserve both the original and translated audio, backed by venture capital from Reddit co-founder Alexis Ohanian.