LetsMT technology

Tilde MT systems are powered by our patented LetsMT technology. LetsMT was developed by a consortium of leading language researchers from across Europe.

Using LetsMT technology, custom MT systems are trained on the customer’s own data, or with the wealth of data available in our multilingual data library. The platform also allows for the addition of translation memories (TMs), glossaries, and monolingual data – significantly boosting the quality and accuracy of translations.

Tilde machine translation lets mt technology

The story of LetsMT

LetsMT was developed by Tilde in cooperation with leading language researchers from the University of Edinburgh, Sheffield University, the University of Copenhagen, the University of Zagreb, and Uppsala University.

Main features:

  • Easy-to-use interface
  • Huge data repository
  • Terminology integration
  • Tag processing
  • CAT tool integration

Users of LetsMT can also take advantage of the Tilde Data Library, one of the world's largest repositories of multilingual data. The Data Library includes over 2.5 billion parallel sentences and 4 million terms in over 100 languages.

Tilde Data Library also includes sophisticated linguistic components, such as named entity recognizers, morphology analyzers, part-of-speech taggers, and other powerful tools for many languages.

LetsMT technology's features include its repository of multilingual data in different domains; correct processing of tags and placeholders, including HTML code; and the processing of many different file formats (including DOCX, TMX, TXT, XLIFF).