Strings are pieces of displayable text, like a headline or button copy. Keys are the IDs used to reference strings in the code. String localization is time-saving for developers, removing any copy & paste work.

Unlike text segments in documents, software strings do not have an order of precedence. This means that there is a slight difference in translation memory (TM) usage. While segments in documents can have 101% TM matches, this doesn’t exist in string-based localization. (101% TM matches mean that not only the copy is matching exactly, but also the previous and subsequent segments match)

Another aspect is context. For string translation, visual context is very important. The strings are usually short and there is no context from previous and subsequent segments.

Phrase Strings offers out-of-the-box integrations with code repositories like GitLab, Github, and Bitbucket. New keys created there will be automatically synced to Phrase Strings to be translated. You can also use the API to build a more customized workflow. For designers, we have integrations with Figma and Sketch. They can trigger the translation process as soon as the UI designs are ready, so that translation and coding happens in parallel. If you need an immediate output, you can have new content be translated automatically with translation memory matches and machine translation.