DeepL, the German-based language AI company known for its high-quality translation tools, has unveiled its first voice translation products: DeepL Voice for Meetings and DeepL Voice for Conversations. The announcement marks the company’s entry into real-time speech translation, a step aimed at revolutionising multilingual communication in both virtual and in-person settings.
Speaking at the company’s flagship event, DeepL Dialogues, held in Berlin, CEO and founder Jarek Kutylowski highlighted the significance of the new offering. “Bringing the quality and security that DeepL is known for to real-time speech translation was the next frontier for us. After years of development and a rigorous beta programme, we’re proud to deliver a solution that addresses real-world communication challenges,” he said.
DeepL Voice for Meetings is designed to enhance virtual meetings by providing real-time caption translations. Participants can speak in their preferred language while others receive translated captions, fostering engagement and inclusivity. Meanwhile, DeepL Voice for Conversations facilitates in-person, one-on-one discussions via mobile devices. The tool provides dual viewing modes, allowing each participant to follow translations on a single device.
The tools support spoken translation in 13 languages, including English, German, Japanese, and Spanish, with captions available in 33 languages already supported by DeepL’s text-based translation platform. DeepL plans to expand this list in the future.
Kutylowski addressed the complexity of developing speech-based translation systems, citing challenges such as pronunciation variations and latency. “Real-time speech translation is fundamentally different from written translation. Misinterpretations can occur easily without accurate and timely processing. Our AI models are trained on diverse data sets to account for accents and varying environments, ensuring the precision businesses require.”
Beta users have praised the solutions for their practicality and inclusivity. Christine Aubry, an internationalisation coordinator at Brioche Pasquier, said: “DeepL Voice for Meetings brought our international teams closer together. Unlike other tools that only support one language per meeting, DeepL Voice allows full multilingual collaboration. It’s by far the most complete tool I’ve used.”
The launch caps a year of innovation for DeepL. The company introduced its enhanced Glossary feature, DeepL Write Pro, and a next-generation large language model outperforming GPT-4 and major competitors in translation quality. Additionally, DeepL raised $300 million in funding earlier this year, pushing its valuation to $2 billion.
DeepL’s solutions are already trusted by over 100,000 businesses and government organisations worldwide. The company aims to extend its impact by breaking down barriers in global communication through the power of AI-driven innovation.
DeepL Voice is now available globally, providing businesses with a secure and efficient way to overcome language challenges in their daily operations.
