DeepL launches voice-to-voice translation and a developer API

In short: DeepL, best known for translating written text, has launched new tools that translate spoken conversations in real time.

What happened

DeepL released a voice-to-voice translation suite that is meant for meetings, mobile and web conversations, and group conversations, such as training sessions. In group settings, people can join using a QR code (a scannable square code that opens a link on your phone).

DeepL is also launching add-ons for Zoom and Microsoft Teams. The company says listeners can either hear the translated audio while someone speaks in another language, or read translated text on screen. These add-ons are in early access, and DeepL is asking organizations to join a waitlist.

For developers and businesses, DeepL is releasing an API, which is a way for software to connect to DeepL’s translation system (like a plug that lets one app use another service). DeepL says this could help companies build custom setups, including for call centers.

DeepL said its current system works in steps, speech to text, translation, then text back to speech. The CEO, Jarek Kutylowski, told TechCrunch that one key challenge is balancing speed and accuracy, meaning reducing the delay between someone speaking and the translation playing back without making mistakes.

Why it matters

If these tools work well, they could make it easier for people who speak different languages to talk during work calls and in-person conversations, without needing a human interpreter. It could also help businesses offer support in more languages, especially where hiring bilingual staff is difficult or expensive.

Source: TechCrunch AI