DeepL becomes real-time translator

With "Voice-to-Voice", DeepL is introducing a new product line for real-time voice translation, intended to assist in meetings, among other things.

listen Print view
Woman with headphones speaks at computer

DeepL is expanding its offering.

(Image: fizkes / Shutterstock.com)

3 min. read

Cologne-based AI speech technology company DeepL launched “the next generation of its translation platform” on Thursday. The new product line “DeepL Voice-to-Voice” is to be understood as an alternative to Google's offering, among other things, as DeepL's new offering also enables real-time speech-to-speech translation. It is intended to be used in virtual meetings, personal conversations, and in “customer-oriented touchpoints via API.” This is intended to enable teams worldwide to “collaborate effortlessly across language barriers,” according to the company.

With DeepL Voice-to-Voice, users are intended to speak naturally in their language, while what is said is to be precisely translated for the recipient. “We combine first-class language models with our proven AI translation technology, which we are continuously developing,” says DeepL CEO Jarek Kutylowski.

According to DeepL, the new real-time translation platform comprises five products:

  • Voice for Meetings: Enables real-time translation on platforms such as Microsoft Teams and Zoom. The early access program starts in June, while registration is possible immediately.
  • The Voice for Conversations function is available immediately. It can be accessed via mobile devices as well as on the web. It can be used where “installing apps is impractical or permissible,” it states.
  • The Group Conversations feature is intended to facilitate multilingual exchange in training sessions, coaching, and workshops. Participants receive a QR code via which they can join directly. With the help of “multi-device access, participants can receive real-time voice translations simultaneously”. The function is scheduled to be available from April 30.
  • Using a voice-to-voice API, companies can integrate DeepL voice translation directly into their own internal applications and customer-oriented tools. The company does not provide a launch date, but interested companies can already register.
  • In addition, DeepL offers optimization functions “with which specific terminology such as industry terms, product names, or proper names can be captured and translated precisely, even with fast or technical language.” This feature is scheduled to be available from May 7.

Videos by heise

At the launch of the offering, DeepL Voice supports all 24 official EU languages as well as Vietnamese, Thai, Arabic, Norwegian, Hebrew, Bengali, and Tagalog. According to independent blind tests by Slator, 96 percent of linguists preferred DeepL Voice over the native translation solutions from Google, Microsoft, and Zoom. In total, DeepL Voice thus covers over 40 languages, according to the company.

In an interview with TechCrunch, Kutylowski said that AI will fundamentally change the customer service experience in the coming years. He further explained that a “translation layer” helps companies offer support in languages where qualified personnel are scarce and expensive to hire.

Most recently, in September 2025, DeepL introduced an AI agent intended to support companies in various areas. The company cited finance, sales, marketing, customer support, and localization as potential areas of application. As an example, DeepL mentioned summarizing findings from a sales team but also automating invoices in the finance department. The company is also apparently considering going public.

(afl)

Don't miss any news – follow us on Facebook, LinkedIn or Mastodon.

This article was originally published in German. It was translated with technical assistance and editorially reviewed before publication.