Breadcrumbs

Exporting and Importing Content in XLIFF for Offline Work

XLIFF is the bilingual format to use when you want to translate or review a job outside the Editor and bring your work back afterwards. The parent article covers the Export and Import steps; this page explains what an XLIFF file actually carries: the pre-translated content, the match origins and suggestions, and the translation status that travels out and back.

Exporting content in XLIFF

An exported XLIFF holds the source and one target language, plus, for each segment, where any pre-translation came from and a translation status.

Pre-translated content and match origins

Before translation begins, Wordbee can pre-fill segments from several sources, and the exported file records the origin of each:

  • Translation memory (reference material): exact matches (identical source), in-context matches (identical text in the same context), and fuzzy matches (close but not identical).

  • Previous version: text reused from an earlier version of the same document, in context or exact.

  • Machine translation: a suggestion from an MT engine.

  • Termbase: terminology suggestions for individual terms.

Memory and terminology suggestions

Depending on the export template you choose, the file also includes memory and terminology hits as alternative suggestions next to each segment, so an offline translator keeps the context.

The richer templates (for example, comments and TM/terminology hits) add these suggestions down to a similarity threshold of 80%. Simpler templates export source and target only.

Status written on export

Wordbee records each pre-translated segment's origin and match strength using the XLIFF state and state-qualifier attributes:

Origin

Match

state

state-qualifier

In-context match (TM or previous version)

context

translated

exact-match

100% match, not in context

100%

needs-review-translation

exact-match

Fuzzy match

fuzzy

leveraged-mt

fuzzy-match

Machine translation

MT

leveraged-mt

mt-suggestion

Auto-correction / rule-based

—

needs-translation

—

Not pre-translated

—

new

—

Importing XLIFF back

When you import the file back, each segment's status is set from its target state, together with the state-qualifier for exact matches. Wordbee has three statuses: Gray (Neutral), Green (Confirmed), and Red (Error, needs attention).

XLIFF target state

Segment status

Color

final, signed-off

Confirmed

Green

translated, new, needs-translation

Neutral

Gray

needs-review-translation with state-qualifier="exact-match" (100% pre-translation)

Neutral

Gray

needs-review-translation (any other qualifier)

Error

Red

needs-review-l10n, needs-review-adaption

Error

Red

needs-l10n, needs-adaption

Error

Red

No state, or any other value

Neutral

Gray

Reading the two tables together explains how pre-translations return. A 100% match that is not in context is exported as needs-review-translation with state-qualifier="exact-match", and on re-import it returns as Neutral (gray), the same as in-context matches (exported as translated). Wordbee treats an untouched exact (100%) pre-translation as neutral rather than an error, and the online QA check still flags any genuine issues. A needs-review-translation state without the exact-match qualifier still returns as Error (red).

This applies when you import a bilingual XLIFF back into a Job, Project, or Resource. Uploading an XLIFF as a new source document is a different process and does not use this mapping.

Learn More