Welcome to the new DeepL API experience: More capabilities, more control and more ways of building with DeepL

At DeepL Spring Launch we revealed how we’ll make the underlying capabilities behind our latest generation of products available to builders and developers via our DeepL APIs.

That’s true. And it’s exciting. But it’s just the start of the story.  

Because our updates to DeepL API don’t just add capabilities. They fundamentally shift how you can work with those capabilities. You can now build transformative new experiences, fine-tune translation and control quality from within your code, and run translation as infrastructure with automatic updates, cost analysis and more.

There’s never been a more exciting time to build with the DeepL API. Here are four reasons why.

Build newly possible experiences with Voice API for real-time speech translation

We’ve launched DeepL Voice API, and it transforms the scope of what you can build. All of those situations and scenarios that depend on real-time, spontaneous, spoken human interaction? You can remove language barriers from each with real-time speech translation. Every capability that’s now available in DeepL Voice core products is also available in the API. What’s more, many of our latest breakthrough features are coming to the API first.

For your organization and users, that could mean customer support platforms that allow any agent to respond to any call. Events where attendees hear every keynote speaker in the language that works for them. Global multiplayer games where users can collaborate, chat and socialize with teammates from across countries and cultures. Frontline clinics and hospitals where every patient can make themselves understood, and cross-language confusion no longer delays diagnosis.

Voice API transcribes and translates speech into up to five different languages simultaneously, from a choice of 45 languages in total. Speech-to-Text is generally available for Voice API now, and you can apply for access to voice-to-voice translation through our Beta program.

How it works:

Voice API configures sessions via a standard REST call before opening a WebSocket channel to stream audio. This enables source transcriptions and tentative translations to display as the speaker talks, with translations then refined as required when sentences are finalized. Glossary support for DeepL Voice API helps to ensure accuracy and precision for specialized conversations in healthcare, legal and more.

Control tone, terminology and structure from within your own code

You’re launching a new customer support assistant in seven markets, but you know that it won’t engage effectively if it adopts the same tone in all of them. Your brand voice is warm in English, formal in Japanese. Your marketing team doesn’t want the same tone for refund queries as for onboarding assistance. Amid the variations though, your legal team needs high-stakes phrases to be translated consistently every time.

With new API access levels to Style Rules and Translation Memory, all of this can be controlled and fine-tuned from within your code. Improved API tag handling means that you can equally control how translated content appears. Language expansion in translation no longer causes headlines to overflow and layouts to shift. Your translations look, feel and engage exactly how they should.

How it works:

DeepL API now has full Create, Read, Update and Delete (CRUD) access to style rules within DeepL. This means you can apply predefined conventions for tone, formatting and grammar at scale, whenever and wherever you need. You can also define your own style rules using natural language prompts, with your logic specifying when they’re applied. Version tracking gives full visibility into when rules are applied, for audit purposes.

DeepL API calls can equally access stored translation memories, referencing them by ID, to leverage previously approved translations. Your code sets the fuzzy match threshold to define when and how they’re applied. In addition, improved tag handling in the DeepL API now sets the standard for preserving HTML and XML formatting, cutting out the time spent checking for distortions and cleaning them up manually. 

Automatically update drop-downs, user choices and features without a developer getting involved

DeepL offers over 100 languages, with new features and capabilities for them launching all the time. You want to keep customers updated about the language options they have within your solutions, but that’s a lot of drop-downs and ‘available feature’ sections to update. DeepL API’s new languages endpoint enables you to programmatically update your applications with the languages available, and the capabilities that can be used with them.

How it works:

Through the endpoint, your code has visibility into language pairs and all features (such as glossaries and style rules) available for each. There’s no more manual management of static feature lists.

Run translation like infrastructure with a full API dashboard

Your FinOps colleagues need to forecast spend for all of your translation workflows, and see how usage and ROI stack up across text, voice and document translation. You don’t have the time to export separate spreadsheets and then work on connecting the dots. With the new usage tab in the API subscription portal, you won’t need to. You can drill down into consumption trends, current and forecasted costs, and full reporting options, all in one place, in real time.

How it works:

You can organize your API usage data by date and usage type, including text, documents and voice minutes, without needing to export to spreadsheets first.

Watch this space for more

Learn more about all our updates and access full documentation at https://developers.deepl.com/docs/getting-started/quickstart. We’ll update the page with new product announcements and API capabilities as they launch.