◇ TMX deduplication online

Deduplicate a TMX file online

Duplicate translation units are the single biggest source of TM bloat and unreliable matches. Upload any TMX and collapse each group of duplicates to exactly one unit — the right one.

━━ the problem

which duplicate wins is usually arbitrary

Most tools keep 'whatever they saw first' or the longest string. Neither knows which translation is actually current or trusted, so your 100% matches are a lottery among stale and fresh variants.

case and whitespace hide real duplicates

"Save settings" and "save settings" look different byte-for-byte but are the same segment. Unicode-normalisation, HTML-entity decoding and whitespace collapse surface duplicates that a naive dedup misses.

you can't see what you're about to remove

Deleting duplicates blind is how translators lose a version they wanted. A preview that shows each group's variants with counts and dates — before you commit — turns dedup from scary into safe.

━━ how TM Cleaner handles it

Pick a winner strategy (most-recent changedate, most-popular, or priority authors), preview every duplicate group with its variants, override any group's winner by hand, then download the deduplicated TMX plus a removed.tmx of everything cut.

━━ frequently asked

how do you decide which duplicate to keep?+

you decide. the default keeps the most-popular variant (frequency), but you can switch to most-recent (latest changedate), rank trusted translators whose variants win, or override any single group by hand in the preview.

does deduplication change my kept segments' text?+

no. dedup only removes whole duplicate units; the text of the unit it keeps is byte-for-byte your original. it never edits translations.

what file formats can I deduplicate?+

TMX (the standard translation-memory exchange format) today. export your TM to TMX from any CAT tool — Trados, memoQ, Phrase, XTM, Smartcat, OmegaT — and upload it here.

━━ also clean

clean your tmx now →

free tier · pay-as-you-go after · no subscription · see pricing