You can search for translations across the entire database. There are filters by language pairs, subjects, organizations, and uploaded file names. This is the primary function for translators, editors, terminologists and other language professionals.
Registration: not required
Access: free of charge
The TM repository is designed for uploading, storing, and using shared domain-specific TMs. The content is available on the basis of mutual benefit: to use the TMs provided by other users, you must upload your TM to the system.
Access: free of charge
A private TM is available only to the user who uploaded it to the repository. To compile a project TM from all suitable records in the TM repository, just upload a set of files you want to translate. The system will automatically fetch 100% and fuzzy matches from the repository and return the project TM as a .tmx file.
Access: paid by subscription
For companies, we offer a translation memory server with advanced capabilities for working with TM, database management and user administration - Memose.
Of all types of TM, the most valuable are domain-specific TMs. In practice, they are not publicly available. Combining the efforts of qualified translators allows them to jointly create high-quality domain-specific TMs, accessible to all members of the product community.
When uploaded to the TM repository, each translation segment is cleared of "technical" metadata - internal hypertext markup tags and placeholders. This way you get the most "clean" linguistic assets that have the following advantages:
Ease of translation and search. Text saturated with internal tags significantly slows down the translation. "Clean" text increases the translator's productivity. In addition, recent advances in automatic tagging make storing tags in a TM completely meaningless.
Independence from CAT systems and file formats. Text without tags can be used in any CAT system and for translating files of any format. You do not need to spend time manually deleting inappropriate tags when substituting a match from a TM.
Suitability for neural machine translation training. The pre-cleaned bilingual TM can be used to train machine translation systems. No need to spend your time and money on cleaning TM every time you use it to train MT engine.
The domain-specific TM accumulates another valuable resource – terminology. In addition to manual search for terms, the repository allows you to extract terminology from TM and create domain-specific glossaries. As a terminology extraction tool, we offer Prospector - a next generation solution featuring high efficiency and low noise. To store and share glossaries, we offer a separate public repository for glossaries - Omnitran.