Phrase NextMT offers enterprise-scale translation in three distinct variants to meet your unique machine translation needs.

Phrase Next GenMT:
Smarter, context-aware translation

Next GenMT phrase

Our most advanced MT solution yet—translating entire text blocks for a more cohesive, natural result.

  • Improved Contextual Accuracy – By considering the broader context, the engine can better understand meaning, reducing mistranslations and ambiguities.
  • Enhanced Fluency – Phrase Next GenMT understands how words relate across sentences, making translations sound more coherent, natural and connected. Longer-range dependencies allow for more natural-sounding translations with improved coherence between sentences.
  • Greater Consistency – Ensures that key terms, tone, and wording remain consistent throughout a document, reinforcing your brand’s voice.
  • Reduced Post-Editing Effort – Produces more accurate and natural translations from the start, so linguists spend less time fixing mistakes.
  • Better Handling of Complex Structures – Context-aware translation ensures related ideas are connected across sentences , improving readability and accuracy, especially in technical or structured content.

How it works

While traditional MT engines deliver fast translations, they often process sentences individually, which can make it harder to maintain consistency. Multi-Segment Phrase Next GenMT takes a more advanced approach, translating entire text blocks together to ensure natural flow, accurate terminology, and cohesive style.

Before (Traditional MT)
Translates individual sentences efficiently
May require additional review to ensure consistency in terminology, formality, and pronoun usage
Delivers fast results with AI-powered fluency and accuracy
Works best with structured content and clear sentence boundaries
After (Phrase Next GenMT)
Translates entire text blocks together for improved coherence
Ensures consistent terminology, formality, and pronoun usage across content
Further reduces manual intervention with advanced AI-driven fluency and accuracy
Speeds up translation workflows by minimizing post-editing
Retains full context to produce more natural, brand-aligned translations

Supercharge your machine translation

With Phrase NextMT, your machine translation doesn’t have to start from scratch. Unlike other generic engines, Phrase NextMT is capable of adapting your translation memories to produce higher quality translations. Wherever a partial match is found, Phrase NextMT will automatically complete the missing content.

Speak your language

Phrase NextMT’s advanced glossary support ensures that your preferred terminology is always used. The simple “search and replace” functionality used by many engines is a thing of the past—with Phrase NextMT, all your terms will be used with the correct morphological inflection, reducing post-editing effort.

Machine translation in style

Many engines struggle with preserving formatting and placeholder tags. Phrase NextMT features advanced tag placement, to ensure that your translations look exactly the way you need them to.

Selecting different MT engines visual | Phrase

Activate with 1 click

Phrase NextMT is part of the Phrase Language AI set of capabilities. Together with advanced MT features like MT autoselect and Quality Performance Score, using MT has never been easier. With a flexible pricing model based on MTUs (Machine Translation Units), you only pay for what you use, gaining access to high-quality machine translation for post-editing from NextMT and other leading engines.

Go further with custom machine translation

Phrase NextMT supports full customization with our innovative Phrase Custom AI. Create an AI language model that speaks your language and redefines machine translation quality.

Phrase Custom AI LP Customization | Phrase

Stay updated

Curious about what’s next for our Generative AI engine? Learn more about upcoming features like document-level context, integrated style guide support, and advanced transcreation.

Prefer a conversation?

Phrase NextMT by the numbers

Increase in Translation Quality
through Translation Memory Adaptation

Increase in Translation Quality
with Morphological Adaption of Glossaries

As a language service provider, we rely on machine translation to help us deliver fast and cost-effective translation for a wide range of customers. With Phrase NextMT, our customer’s translation memories no longer compete with machine translation, but actively improve its quality. Phrase NextMT’s advanced glossary functionality is especially useful for ensuring that our customer’s preferred terminology is always respected.

Václav Baláček

Director at Czech Translations

With the addition of Phrase NextMT to Phrase Language AI, Phrase has a complete MT feature set that responds to every market requirement, from AI-driven engine selection and instant quality estimation to easy customization and glossaries.

Jourik Ciesielski

MT Expert at Nimdzi

Here to field your questions on Phrase NextMT

Which language pairs does Phrase NextMT support?

Phrase NextMT currently supports the following language pairs in both directions:

  • English-Czech
  • English-Spanish
  • English-French
  • English-Russian
  • English-German
  • English-Italian
  • English-Dutch
  • English-Chinese (Simplified)
  • English-Swedish
  • English-Japanese
  • English-Portuguese
  • English-Korean
  • English-Polish
  • English-Danish
  • English-Greek
  • English-Norwegian
  • English-Indonesian
  • English-Hungarian
  • English-Finish
  • English-Slovak
  • English-Vietnamese
  • Spanish-Catalan

We will continue to add new languages to Phrase NextMT.

Does Phrase’s MT autoselect prioritize Phrase NextMT?

The MT autoselect algorithm available in Phrase Language AI is impartial when it comes to recommending the optimal engine for a given language pair and content type.

Its recommendation is based on past performance of engines. MT autoselect will not prioritize Phrase NextMT over other engines already available.

Was customer data used to train Phrase NextMT?

Phrase NextMT was developed using commercial and public datasets. No customer data was used to train the engine.

Want to find out more?