What a TM contains
A translation memory holds entries (segments). Each entry has:- A source text (e.g. “Add to Bag”, “Your order has shipped”).
- An optional context to disambiguate identical source strings.
- A translation per language.
How TM helps while translating
Once a TM applies to a project, it works automatically in the editor:- When you edit a value, TextSetu compares the source text to TM entries.
- If a stored entry matches at ≥75% (exact or fuzzy), it’s offered in the TM suggestion panel.
- Click a suggestion to drop the stored translation into the cell — then tweak if needed.
Managing a TM
A translation memory has its own workspace:
Secondary areas (Import, Export, Languages, Activity, Access, Settings, Applicable
projects) live under a More menu in the TM’s tab strip.
In Settings, the General and Matching cards each have their own Save
button, while the Workflow toggle saves automatically; turning approval off
discards any in-progress entry proposals awaiting review.
On the Entries list, the Filters panel narrows by missing-in-language and
approval status, with removable chips.
Each entry keeps its own history of changes.
Importing and exporting
- CSV — columns for
source_text,context, then one column per language. - TMX — the industry-standard Translation Memory eXchange format (TMX 1.4), for interoperability with other tools. Available for both import and export.
Keeping the TM in sync
When a source string changes, TextSetu can prompt you about affected TM entries so your memory stays accurate rather than suggesting stale translations.TM vs. Glossary — what’s the difference?
Use both together: the glossary keeps individual terms right, the TM reuses entire
sentences.
Related: Glossary · Translating · Import & Export

