BabelBee Supports XLIFF 2.0/2.1: Fewer Steps Between CAT and AI

For translation project managers, bringing AI into an established workflow can involve more file handling than expected. Export from one CAT tool, convert in another, run the translation, then reverse the process. Each extra conversion means another check of segment boundaries, tags and existing translations.

BabelBee now supports XLIFF 2.0 and the core content of XLIFF 2.1, while retaining the existing XLIFF 1.2 workflow. Upload an .xlf or .xliff file, use BabelBee for AI translation, review and human editing, then export in the original version. There is no need to downgrade a supported 2.0 file to 1.2, or change a workflow that already uses 1.2.

This update also adds native round-trip support for TXML, SDLXLIFF and MQXLIFF. Files within the supported scope can follow a shorter route:

CAT bilingual file → BabelBee translation and review → Same-format bilingual file → Original CAT tool for review and delivery

For example, a team working with Wordfast TXML can upload TXML directly and download TXML after translation. This can remove the need to pass the file through memoQ and a separate converter just to use AI. SDLXLIFF and MQXLIFF have their own native processing paths too. The original CAT project still produces the final Word document or other deliverable.

For these three native CAT formats, BabelBee translates only unlocked, editable segments with empty targets by default. Existing translations and confirmed content remain intact. Original segment boundaries, supported formatting tags, comments and history stay with the file. New SDLXLIFF translations are marked as drafts, and new MQXLIFF translations as machine translated, so reviewers can continue their work. Missing or reordered tags trigger an export error so they can be corrected.

When a different XLIFF version is required, the toolkit also offers conversion in both directions between 1.2, 2.0 and 2.1. This currently covers plain-text core XLIFF with one segment per unit, excluding inline tags, groups and proprietary metadata. For formatted translation projects, use the original-version round trip where possible. Upload native TXML, SDLXLIFF and MQXLIFF files through the translation project entry point.

Native CAT support currently covers a defined set of file structures. It does not provide arbitrary conversion between these three formats, or support TXLF and complete project packages. Start with a small file and check the full journey through re-import and final document generation. We will continue refining compatibility with feedback from PMs using real project files, helping teams spend less time on conversion while keeping their familiar CAT tools.