LinguaAI Toolkit

Description

CONVERT FOR TRANSLATION, THEN BACK Word documents, PowerPoint decks, Excel workbooks, subtitle files (.srt) and XML become XLIFF for translation in any CAT tool, then rebuild into the original format when the translation is done. Formatting, tables, footnotes, speaker notes and in-cue line breaks survive the round trip because the original file travels inside the XLIFF and the translation gets written back into it. Only confirmed segments ship their translation, so a half-finished file looks visibly half-finished instead of passing as done. Convert a whole job at once, and drop the translated XLIFFs back on the app to rebuild them all after one folder pick. Say you've got a spreadsheet with five columns and only one needs translating: pick the source column and the translation column, and the rest never enter the XLIFF. For a Word table whose left column holds the client's own labels, one tick locks that column so it comes back untouched. HANDLE AGENCY PACKAGES WITHOUT THE AGENCY'S TOOL Unpack project packages (.sdlppx) to translate their files anywhere, then build the return package (.sdlrpx) the agency expects, with only the translation swapped in. Same goes for .mqxlz packages and the .glp packages GlobalLink agencies send, whose buried client instructions get surfaced right where the work starts. Embedded memories convert to TMX and the integrated termbase exports along the way. TAME YOUR MEMORIES AND TERMBASES Convert .sdltm memories to standard TMX. Merge TMX files. Split a client TM that mixes en-US, en-GB and bare en into one clean file per language pair. Build a TM from any two-column bilingual spreadsheet or CSV. Convert termbases (.tbx, .sdltb) to CSV and ready-to-import glossaries, term statuses included, so a term the client rejected shows up labelled as rejected instead of posing as one more valid option. EXTRACT AND ALIGN WITH AI, ON YOUR OWN KEYS Pull terminology and style guides out of your translation memories or source documents, with confidence scores, frequencies and definitions. Pair a document and its translation into a TMX with Document Alignment: the model only decides which sentence goes with which, the text itself is copied straight from your files, and sentences split or merged in translation get marked for you rather than glued together. Convert PDFs and scans to translatable Word documents through OCR that follows reading order, not pixel position. Bring your own key for Anthropic, OpenAI, Google, Mistral or Perplexity, whose router also serves Claude and other models on one Perplexity key. Nothing routes through anyone else's account. PRIVACY FIRST Your API keys stay in your Keychain and never leave your Mac. File processing happens on your Mac; AI requests go straight to the provider you chose. No analytics, no tracking, no data collection, no subscription. Pairs with LinguaTrans, the suite's CAT editor: converted files hand off with one click, and finished translations convert back.

A package's memories merged with their mark-downs, and Claude Sonnet 5.5. - Project Packages: after unpacking, one button merges the package's memories into a single TMX, and every unit from a memory the package marks down carries that mark-down inside the file. LinguaTrans reads it at import, so the client's scores come out right without setting anything by hand; other tools ignore it. The unpack result stays on the tab while you visit other tabs. - TMX Merge: a mark-down per file, kept inside the merged TMX on each unit, for the same purpose. - Claude Sonnet 5.5 is in the model list, above Sonnet 5, at the same price. - About's copyright line keeps its year current.

Information

Seller

Richard Jacobs

Category

Productivity

Version

1.16

Requires

iOS 26.0+

Size

6.0 MB

Age Rating

4+