5 key takeaways from DeepL Spring Launch

In this post
- We launched a new way of working with language
- Real-time, voice-to-voice translation for every business communication
- AI for language that’s designed to work with language experts
- Superior AI translation quality that’s preferred by language experts
- Build instant translation into your products and platforms with DeepL API
- Turning language from barrier to business advantage
When you’re part of product development, there’s nothing more exciting than introducing your work to the world. I’ve never looked forward to a launch more than I have to today’s — and DeepL Spring Launch didn’t disappoint.
We didn’t just share a new set of tools or features with you at our virtual event. We unveiled a blueprint for how borderless businesses can operate in the AI era. And we shared our commitment to helping you get there through end-to-end, enterprise-grade products that represent a wholly new way of working with language.
Here are 5 key takeaways from DeepL Spring Launch:
We launched a new way of working with language
DeepL is now the unified operating layer for language. We’ve built a multilingual AI platform that applies highly accurate, customized translations to every workflow that matters, at the speed global business requires, and with the control that enterprise needs.
The new DeepL Translator platform offers a new way of working with language and an alternative to TMS and LSPs that combines speed, control and transparency with superior translation quality. Teams can now generate accurate, compliant and on-brand translations from within any tool or system they use, and run entire translation flows in minutes! There’s no context-switching, no toggle tax, and no need to pause projects while you brief in translation separately.
Real-time, voice-to-voice translation for every business communication
Voice-to-voice translation is now live, communicating meaning across languages, in real time. It’s the biggest transformation to how colleagues, partners and customers collaborate that we've ever launched. It enables instant understanding, natural conversation flow and inclusive experiences. It does this across virtual meetings, in-person and group meetings, customer support calls, and every other conversation that matters. Global business will never be the same.
AI for language that’s designed to work with language experts
We’ve designed DeepL Translator to work with your language experts and localization teams:
- Translation Flow saves huge swathes of their time by taking content directly from the systems where it lives and orchestrating translation workflows end-to-end.
- DeepL quality assessments loop in human review when it’s required, and avoid the need to painstakingly review all translated text.
- Translation Flow’s professional editing experience takes reviewers to exactly where their input is needed, so they can edit and update translations in real-time.
- Customization Hub scales the impact of their knowledge and the choices they make on translation.
By turning language into an operating capability and growth engine, we’re freeing language experts from repetitive, time-consuming tasks, and empowering them to shift to a more strategic role.
Superior AI translation quality that’s preferred by language experts
All of this is underpinned by a foundation of accuracy and precision. Take real-time voice translation, for example. This month, Slator released a market assessment of AI translated captions for real-time voice translation in which DeepL Voice outperforms all rivals by a huge margin on both the quality and stability of translations: the two elements that are essential for effective voice-to-voice translation. DeepL Voice was preferred by 96% of Slator’s human experts over translations from Zoom, Microsoft and Google Translate.

In addition, since launching our new model last month, we’ve commissioned blind evaluations with professional linguists to make sure we’re continuously raising the bar on quality and trust, with no scope for bias.
In our latest round of tests, we evaluated results across 16 priority language pairs. DeepL’s new model won 94% of those tests against major AI competitors:
• 100% of language pairs vs Google Translate
• 100% vs ChatGPT-5.2 (high reasoning mode)
• 100% vs Microsoft Translate
• 88% vs Google Gemini 3 Pro (thinking mode)
• 81% vs Claude Opus-4.6
At the end of the day, translation isn’t just about words. It’s about meaning, precision and trust. That’s what we’ve tested for, and what we continuously aim to deliver to our users.
Build instant translation into your products and platforms with DeepL API
We’re making our leading features and capabilities available in DeepL API and DeepL Voice API. These empower your builders and developers to embed groundbreaking AI for language into your own products, platforms and workflows. Whether it’s Articulate enabling authors to create course content in multiple languages, or Amazon Connect bringing real-time voice translation to customer support platforms, our APIs are helping to apply language as a growth engine, everywhere.
Turning language from barrier to business advantage
In the US alone, nearly 70% of businesses report daily disruption from language barriers, while 60% say it has delayed global expansion. The capabilities we showcased this week are ready to change that. They are a growth engine, designed to turn language from a barrier into your business advantage. And it’s all available to you now.
Catch up on all the highlights from Spring Launch, and get started with the operating layer for language. Everything you need is on the DeepL Spring Launch Hub.