DeepL is rolling out DeepL Voice-to-Voice which is a real-time speech translation system aimed at breaking language barriers in communication whether through digital means or face-to-face. This will be the first speech-to-speech translation service by the company, and it will allow people to translate the spoken conversation live on the spot during a meeting, a customer visit, or any other type of communication through the API integration of the product.
Jarek Kutylowski, Founder & CEO of DeepL said: “Today, we reach another frontier in translation: real-time, spoken communication. Our mission has always been to break down language barriers and we’ve now overcome one of the biggest of all. DeepL Voice-to-Voice allows everyone to speak naturally in their own language without the friction or cost of interpreters. We’re fusing world-class voice models with the gold-standard translation AI we’ve been pushing to new heights. Now, expertise is all that counts, not language.”
Expanding Real-Time Communication with DeepL Voice
DeepL Voice has been developed to tackle one of the final significant obstacles in enterprise communicationcorrect, instantaneous spoken translation. The recently launched Voice-to-Voice package features a number of functionalities specifically designed for contemporary business requirements:
- Voice for Meetings: Makes it possible for instant translation to take place on communication software like Microsoft Teams and Zoom. This way, everyone attending the meeting can speak in their respective languages without any language barriers. (Early access starts in June.)
- Voice for Conversations (Mobile & Web): Extends beyond mobile to support browser-based interactions, making it ideal for environments where app installations are restricted. (Now generally available.)
- Group Conversations: Supports multilingual collaboration in training sessions and workshops, allowing participants to join via QR code and receive simultaneous voice translations. (Available from April 30.)
- Voice-to-Voice API: Allows businesses to embed DeepL’s real-time voice translation into internal tools and customer-facing applications. (Early access ongoing.)
- Customization with Spoken Terms: Enhances translation accuracy by recognizing industry-specific terminology, product names, and technical language—even in fast-paced conversations. Glossary integration ensures consistency across use cases. (Available May 7.)
To broaden accessibility, DeepL has also introduced a self-service model, enabling smaller teams to purchase and deploy Voice solutions directly online, including free trial options.
Also Read: Moloco Launches AI-Powered Performance CTV to Help App Marketers Drive Measurable Growth
Broad Language Support and Performance Validation
DeepL Voice has expanded its language support to more than 40 languages, not just all the official languages of the European Union, but also Vietnamese Thai Arabic, Norwegian Hebrew Bengali, and Tagalog.
Third-party blind tests by Slator have shown excellent performance results, where 96% of the linguists preferred DeepL Voice compared to other solutions from Google, Microsoft, and Zoom. DeepL Voice scored above 96/100 in quality assessments for its Zoom and Microsoft Teams integrations, demonstrating its naturalness of speech and contextual understanding.
Yoichi Okuyama, Head of DX System Department at Pioneer, added:
“Relying solely on English proficiency for global collaboration often slowed us down, as team members hesitated to contribute complex ideas. By implementing DeepL Voice, we’ve removed that friction and created a more inclusive environment where everyone can speak confidently in their native language. This shift has helped accelerate our business processes; with barriers removed, we’ve seen more active participation and faster decision-making across our global teams. It’s transformed translation from a technical necessity into a key enabler for speed and efficiency.”
Advancing the Next-Generation DeepL Translator Platform
Alongside the Voice-to-Voice launch, DeepL is evolving its core product into a next-generation AI-powered Translator platform, aimed at modernizing enterprise translation workflows.
Jarek Kutylowski added: “Global businesses no longer have a translation problem; they have an operating model problem, with today’s language solutions often being too slow to scale and a costly drag on growth for businesses. We’re bringing translation and language fully into the AI age. By centralizing translation operations in an AI-first, multilingual platform, every team can access fast, high-quality translations without being held back by legacy tools or relying on expensive third-party language services.”
Key Platform Enhancements
- Translation Flow: Conveniently blends translation with present workflows, thereby eradicating both the time lags and manual effort translation
- processes require. Translation Quality Assessment: It also concurrently outputs the evaluation parameters so that translation teams can measure the translation’s trustworthiness right away.
- Continuous Improvement: Besides that, it can also identify changes that the users make and utilize them to progressively increase the degree of accuracy, besides that, it is also capable of tailoring itself to each and every organization’s requirements.
The platform, which integrates translation into everyday workflows, allows the same day execution of tasks, significantly increases consistency, and makes the company less dependent on outside services.
Geoffrey Wright, Global Solution Owner – GenAI and Digital Experience at Mondelēz International recently highlighted the impact of this shift:
“At Mondelēz, we don’t settle for slow—on the road or in our workflows. Our old translation process was like driving on a flat tyre, but DeepL is full service at 100 mph. By embedding their Language AI, teams like M&A and Legal are handling sensitive documents with top speed and total confidentiality. When you make the impossible look that easy, word travels fast; we’ve seen adoption accelerate across the entire organization.”