How is translation memory (TM) used for a Transifex Native project and resource?

Translation memory works consistently on any new content added on a Transifex project regardless of the action or method followed to push that content in the project. Specifically, whenever new content is pushed to Transifex, TM will check for exact matches and efficiently populate translations for source strings that have 100% similarity match with already translated source strings.

Since Transifex Native follows the ICU syntax to mark plurals and variables, the exact matching performed will look for similar syntax.

For example, if you have a TM entry containing a variable like Hello %(username)s! that has a French translation, and you add new content through Transifex Native SDK that is Hello {username}, then those two entries will not match due to difference in the way the variables are defined but an 83% similarity match will still be provided by TM as follows:

You can check the table below where more examples can be found: