DeepL, the Cologne‑based language‑AI company best known for its text translation tools, unveiled a full‑scale voice‑to‑voice product suite on Thursday. The new offering, branded DeepL Voice‑to‑Voice, targets live business communication and spans four distinct scenarios: virtual meetings, mobile and web conversations, group settings for frontline workers, and an enterprise API that lets developers embed the technology into their own applications.

The suite already supports more than 40 languages, covering all 24 official European Union languages plus Vietnamese, Thai, Arabic, Norwegian, Hebrew, Bengali and Tagalog. Voice for Conversations, which enables real‑time translation on mobile and web without requiring a separate app, is now generally available. Voice for Meetings, integrated with Microsoft Teams and Zoom, will begin an early‑access program in June, allowing participants to speak in their native tongue while hearing simultaneous translation in the language of their choice.

Developers can also apply for early access to the Voice‑to‑Voice API, which lets businesses embed DeepL’s translation engine into customer‑facing tools such as call‑center software. A customization feature called Spoken Terms, slated for general release on May 7, will let the system learn industry‑specific vocabulary, company names and personal names, further tailoring translations to corporate needs.

CEO and founder Jarek Kutylowski framed the launch as “another frontier in translation,” emphasizing that DeepL Voice‑to‑Voice lets users converse naturally without the friction or expense of human interpreters. The company also stressed its security posture: voice data is never used to train models and is deleted once a call ends, a claim that sets it apart from many consumer‑grade AI voice products and aims to satisfy regulated industries.

Technically, the system follows a three‑step pipeline: speech is converted to text, the text is processed by DeepL’s award‑winning translation engine, and the result is rendered back into speech. DeepL’s competitive edge rests on the middle step; the firm argues its text models outperform rivals, a benefit that carries through to the spoken output.

Independent blind evaluations commissioned by DeepL and conducted by language‑industry researcher Slator found that 96 % of professional linguists preferred DeepL Voice over native translation solutions in Google Meet, Microsoft Teams and Zoom. The suite scored 96.4 out of 100 for Zoom and 96.3 for Teams, with reviewers citing superior fluency and contextual accuracy.

During a live demo at DeepL Connect Seoul, Chief Product Officer Gonzalo Gaiolas highlighted a current limitation: a one‑to‑two‑sentence delay between a speaker’s utterance and the translated output. He attributed the lag to differing word orders across languages, a challenge the company says it will continue to address through model improvements.

Looking ahead, DeepL plans to reduce latency further and introduce a voice‑preservation feature that maintains the speaker’s original vocal characteristics in the translated audio. That capability is expected by the end of 2026.

DeepL enters a crowded market populated by well‑funded competitors. Sanas, backed by Quadrille Capital, focuses on real‑time accent modification for call centers. Dubai‑based Camb.AI targets speech synthesis and translation for media dubbing, while Palabra, supported by Reddit co‑founder Alexis Ohanian’s Seven Seven Six, is developing a voice‑preserving translation engine. Tech giants Google, Microsoft and Zoom already offer meeting translation features, making DeepL’s simultaneous integration with those platforms both a strategic challenge and an opportunity.

By betting on translation quality—a hallmark of its brand—DeepL hopes to carve out a niche among enterprise customers who value accuracy and data security over the distribution advantages held by incumbent platform providers.

Used: News Factory APP - news discovery and automation - ChatGPT for Business