Translation Memory
What translation memory is, what it does well, and why it remains useful even in AI-assisted translation workflows.
Translation memory is a system that stores previously translated segments so they can be reused when the same or similar text appears again. In documentation, product UI, and recurring instructional material, this reduces duplicated effort and improves consistency over time.
Translation memory is often discussed alongside glossaries, but the two are not identical. A glossary defines preferred terms. Translation memory records how larger segments were translated in context. Together they help translators work faster while keeping wording stable across releases.
Even in AI-assisted workflows, translation memory remains useful. Generative tools can produce a fresh draft each time, but that does not guarantee stable terminology or phrasing. Translation memory gives teams a repeatable baseline and helps reviewers see when new output diverges from established usage without a good reason.
This matters most when content is updated continuously. A long-running multilingual project needs more than one-off translation quality. It needs editorial memory. That is why the wiki entry on Translation Memory and Glossaries remains relevant even in a repository that also discusses AI-assisted translation.
For LetBooks, the value is practical. The documentation corpus spans product, accessibility, localization, and metadata topics. Some phrases should stay stable across languages and across time. Translation memory is one of the tools that helps keep that stability visible instead of relying on informal recollection.