Translation memories are part of the customization hub, which groups DeepL's customization features in one place. Other features you can find here are the following.
Translation memories are mostly used by professional translators to store and reuse previously created translations. Use the translation memory feature from DeepL to upload your translation memories as a TMX file and use them for your text and file translations.
For a detailed description about how to use translation memories within DeepL, go to this article.
DeepL for Enterprise users have the option to update translation memories from translated files. Learn more about updating translation memories here.
Available platforms
Translation memories are available on the following platforms:
- DeepL website
- DeepL desktop apps
Available languages
Translation memories are currently available for the following languages:
|
|
Translation memory eXchange files
The translation memory exchange format (TMX for short) is a XML-based file format used for storing translation data. It is mainly used for translations by professional translators in CAT tools.
A TMX file contains multiple text segments and all of their translations. For each language there is an element called translation unit variant <tuv> which contains information about the language and the text segment in a child element <seg>.
Both elements are child elements of a parent element called <tu>.
Example:
<tmx version="1.4">
<header creationtool="CAT Tool" creationtoolversion="1.0" datatype="PlainText" segtype="sentence" adminlang="en-US" srclang="en"></header>
<body>
<tu>
<tuv xml:lang="en">
<seg>Hello World</seg>
</tuv>
<tuv xml:lang="es">
<seg>Hola Mundo</seg>
</tuv>
</tu>
</body>
</tmx>
DeepL team admins can upload a TMX file to share the translation memory with their whole team. TMX files can be retrieved from CAT tools via the respective export options. Consult the support page of your CAT tool provider for further information about the TMX export of your CAT tool.
Types of matches for translation memories
100% match
The contents of original text and the source segement of the translation memory match to 100%.
In-Context Exact match
The contents of original text and the source segement of the translation memory match to 100%. Additionally, the segment before and after also match 100%.
Smart match
The original text and the source segment of the translation memory match, but not at 100%. The translation is automatically adjusted according to the differences.
Limits
There is no limit regarding the number of translation memories that can be uploaded for DeepL translations, however as for all DeepL services there is a fair usage policy in place. Also, only one translation memory can be applied to a translation at a time.
TMX files should have a maximum file size of 1GB for the upload.