TraduQ CAT is a free CAT tool and XLIFF editor in your browser. Translate Trados and memoQ packages, SDLXLIFF, MQXLIFF, MQXLZ, XLIFF and Wordfast TXML files, Word (.docx), Excel (.xlsx), PowerPoint (.pptx) and InDesign (Beta) files, PDFs (text only, delivered as Word) and software and text files (JSON, PO, .properties, .resx, Apple .strings and plain text, plus Markdown, subtitles, YAML and Qt) with TM matches, a termbase and QA, and download it in its own format. No install, nothing uploaded.
Your files are read in your browser and never uploaded. Your work is saved only in this browser.
Choose a Word document (.docx), an Excel workbook (.xlsx), a PowerPoint presentation (.pptx), an InDesign file (.idml, Beta), a PDF (text only), an SDLXLIFF, MQXLIFF, MQXLZ, XLIFF or XLF file, a Wordfast .txml file, a Trados package (.sdlppx), a memoQ package (.mqout), or a text or software file (.txt, .md, .srt, .vtt, .po, .pot, .json, .arb, .yaml, .yml, .properties, .resx, .strings, .stringsdict, .xcstrings, Qt .ts), up to 200 MB. Nothing is installed and Trados or memoQ are not needed.
With a file open, the editor takes the whole window: the grid, a results panel with TM matches and terms, and a files panel for packages with several files. It uses the package's TM and your own TMX, SDLTM or glossary. Matches that differ only in numbers, dates or codes are updated from the source, and Pretranslate fills empty segments. In Word, Excel and PowerPoint, bold, italic and underline show as formatting (Ctrl+B/I/U, or Cmd on a Mac). QA and concordance search are built in.
Download your file with the translations, in its own format: only the targets, their statuses and, in some formats, the file's language marker change; everything else stays as it was. A PDF comes back as a Word document (.docx). You can also export the TM with what you confirmed as TMX. From a Trados package you can also export a return package (.sdlrpx) with your translated files; it has been tested in real Trados round trips, but check with the agency the first time.
It opens Trados SDLXLIFF files, memoQ MQXLIFF and MQXLZ files, XLIFF 1.2 and 2.0 files (.xlf, .xliff), Trados project packages (.sdlppx) and memoQ packages (.mqout), up to 200 MB. A memoQ package goes back as the translated MQXLIFF files, not as a .mqback delivery package. It also opens Word documents (.docx), Excel workbooks (.xlsx) and PowerPoint presentations (.pptx) and translates them in the same grid. InDesign files (.idml, Beta) open too; see the InDesign question below. PDF files open as text only: you translate the text and get a Word document back, without the layout. Text and software files open too: plain text (.txt), Markdown (.md, .markdown and the other Markdown extensions), subtitles (SubRip .srt, WebVTT .vtt), gettext PO and POT (.po, .pot), JSON and Flutter ARB (.json, .arb, i18next included), YAML (.yaml, .yml), Java .properties, .NET .resx, Apple .strings, .stringsdict and .xcstrings, and Qt Linguist .ts. Wordfast TXML files (.txml) open as bilingual files; that format is in beta, so check the first one you deliver with whoever sent it. Only the text is translated: keys, comments, placeholders and code stay as they are.
No. The file is opened and edited in your browser and its contents never leave your computer. The only thing we count is that the tool was used, with the file extension and a size range, never the file, its name or your translations.
Yes. A Trados package that includes a translation memory uses it right away, and the matches appear next to each segment. You can also add your own TMX or SDLTM files. Everything you confirm is added to a working memory that you can export as TMX.
Yes. Add a glossary or termbase as CSV, TXT, Excel, TBX or MultiTerm XML (memoQ and Trados termbase exports included) and the terms found in each source segment are listed and can be inserted with a shortcut. Terms that were not used in your translation are flagged by the QA checks.
Tags appear as small chips that you can insert, move or copy with the keyboard. If a confirmed segment has missing, extra or unpaired tags, the final file cannot be downloaded until you fix it, so you do not hand back a file that breaks in Trados or memoQ. Wordfast TXML files are the exception for missing or extra tags: Wordfast accepts them, so they are only warnings. You can still save your progress at any time.
Your translations and their status, and nothing else. The editor writes into the original file instead of rebuilding it, so everything outside the targets, their statuses and, in some formats, the file's language marker (such as a YAML file's root locale key or a Qt file's language) stays byte for byte as it was, and a file you did not edit is downloaded exactly as it came. Packages and containers are rebuilt as zips with the files inside unchanged except your targets and statuses; minimal markup is added only where a target, or a plural form the target language needs, was missing. A PDF is the exception: you get a new Word document built from its text (see below).
No. The editor suggests only what is already in your translation memories and glossary. It never sends text to a translation engine or to any server.
The editor is designed for a computer with a keyboard. On a tablet it works with the side panel under the grid. On a phone it shows the progress and the QA list in read-only mode.
Not entirely. It is a browser CAT tool for bilingual files and packages, Office documents, PDFs and software strings, and you translate without installing anything. For a Trados or memoQ package, the final document is generated by whoever sent it to you, from the file you give back. It translates Word (.docx), Excel (.xlsx) and PowerPoint (.pptx) files directly and returns them translated in the same format, a PDF as text that comes back as a Word document, a Wordfast .txml as a .txml, and text and software files (.txt, .md, .srt, .vtt, .po, .json, .arb, .yaml, .yml, .properties, .resx, .strings, .stringsdict, .xcstrings, Qt .ts) that come back in their own format. Features such as live project settings, plugins and machine translation are not included.
Yes, and only when you opened a Trados project package (.sdlppx). The export is an .sdlrpx file with the translated target SDLXLIFF files and the project file switched to a return package, with no source files, reports or translation memories. It has passed real round trips in Trados, but it is fairly new, so check with the agency the first time. The same tag check as the final download applies: segments you edited must have their tags right, while tag problems already in the file you received are listed and go back unchanged. Untranslated segments can be included, with a warning. A memoQ package (.mqout) goes back as the translated MQXLIFF files (a zip of .mqxlz files), not as a .mqback delivery package: your PM imports them into the memoQ project (Import, or Update from bilingual), which updates the documents already there. We do not certify it with Trados and cannot promise that every agency accepts it.
Yes. When the package inspector opens a Trados project package (.sdlppx) or a memoQ handoff package (.mqout), the Translate in TraduQ CAT button opens it here directly, without choosing the file again. It stays in your browser the whole time.
Yes. When you confirm a segment, its target also goes to every other segment with the same source text and tags that is still empty, and to repetitions filled that way that you have not touched since. Your own drafts and confirmed translations are never overwritten, and Trados files mark those segments as auto-propagated. Spell check uses built-in Hunspell dictionaries for more than 40 languages (one Spanish dictionary for every variant, English US and UK, Portuguese Brazil and Portugal…), checked in your browser: nothing is uploaded. Right-click a word for suggestions or to add it to My dictionary; QA lists spelling issues. A few languages, such as Chinese, Japanese, Korean, Finnish and Malay, use your browser's spell check. You can turn either one off in Editor settings, in the toolbar.
Yes. Open a .docx file and it is split into segments that you translate in the same grid, with translation memory, termbase, QA and tags. The text of the body, tables, headers, footers, footnotes, endnotes and text boxes is translated; text inside images is not. The translated file is a .docx where only the translated text changes and the rest of the document stays as it was, with your target language set on the translated text. Old .doc files and encrypted files are not supported: save the document as .docx first.
Fields such as page numbers, dates and cross-references are kept as tags and are not translated. A table of contents is rebuilt at export from your translated headings, and index entries are rows you translate like any other. When Word opens a file the export changed, it asks to update fields: choose Yes (and Update page numbers only if it asks about the table of contents) so the page numbers and the index match the translation.
Paragraphs with tracked changes are shown read-only and are not written, so accept or reject the changes in Word first if you want them translated. Comments are not translated, but their marks stay where they were.
Yes. The translated text in Word (.docx) and PowerPoint (.pptx) files is marked with the target language you chose when you opened the file, so Word and PowerPoint spell-check it in that language. Untranslated text keeps its own language. Excel files have no per-text language, so nothing is set there. Right-to-left languages such as Arabic and Hebrew, and Chinese, Japanese and Korean, can be used, but paragraph direction (right-to-left) and East Asian font hints (Chinese, Japanese, Korean) are not set: check the layout in Word, Excel or PowerPoint.
It is a Word table (.docx) with one row per segment: number, source, target, status and notes. Send it to a reviewer, who can edit the target column, with or without tracked changes. Then use Import reviewed table… to bring the changes back: you see the list of differences, choose what to apply, and they go in as one undoable step. Changes that break the tags, or touch locked or read-only segments, are listed but not applied. It works for every file type the editor opens, not only Office files. Reviewing in TraduQ itself (Working as: Reviewer) keeps your edits as tracked changes instead, and Changes report (.docx) lists only the segments you changed.
Yes. Open an .xlsx file and the text cells of every sheet become segments that you translate in the same grid, with translation memory, termbase, QA and tags for bold or coloured words. Formulas, numbers, dates and true/false values are never read or changed, so the workbook keeps calculating as before. The same text in several cells is one segment, translated once for all of them. Hidden sheets, rows and columns are included and marked hidden in the grid. The header row of an Excel table (Format as Table) and the label of its totals row are shown locked: formulas refer to the table's columns by name, and Excel expects the totals label to match the table. The translated file is an .xlsx where only the translated text changes.
No, not for now. Sheet names stay as they are, because formulas and charts refer to sheets by name and renaming one would break them. Comments and notes are not translated, and neither is text in charts, shapes and text boxes. Old .xls files are not supported, and neither are password-protected workbooks: save an .xls as .xlsx, or remove the password in Excel (File › Info › Protect Workbook › Encrypt with Password), first.
Yes. When you open an Excel file, choose source and target columns instead of translating in place. For each sheet, pick the source column (for example B), the target column (for example C) and whether row 1 is a header; the languages are suggested when the header says something like English, es or ES-419. Each source cell with text becomes a segment, and its translation goes into the same row of the target column (the cell is created if there is none). A target cell that holds a formula or a number, or is inside merged cells, is shown locked and left as it is. Text already in the target column appears as a draft and is only rewritten if you edit or confirm it. Other columns and sheets are left untouched.
Yes. Open a .pptx file and the text of every slide becomes segments that you translate in the same grid, with translation memory, termbase, QA and tags: slide titles, text boxes and other shapes, shapes inside groups, tables and the speaker notes, in slide order. Hidden slides are included and marked hidden in the grid. Slide numbers and dates are kept as tags. The translated file is a .pptx where only the translated text changes, with your target language set on it so PowerPoint spell-checks it in that language. Old .ppt files are not supported, and neither are password-protected presentations: save the file as .pptx, or remove the password, first.
Slide masters and layouts are left as they are, including the prompt text of their placeholders, such as Click to add title, which does not show in the slide show. Text in charts and SmartArt graphics is not translated, and neither is text inside images. Some shapes also have a backup copy for older PowerPoint versions; when it differs from the main text, that copy keeps its original text.
Usually it does, but text boxes set to Shrink text on overflow need one click. PowerPoint keeps the font size it saved for the original text until the box is edited, so a longer translation can spill out of the box when you open the file. Click in the box, type a space and delete it, and PowerPoint shrinks the text to fit again.
Yes, as text. Open a .pdf file and its text becomes segments that you translate in the same grid, with translation memory, termbase and QA. The layout is not kept: you download a Word document (.docx) with the text, headings and lists in reading order. Headers and footers repeated on every page are taken out of the text, translated once and written into the Word file's header and footer, with page numbers as fields. Up to 2,000 pages. Source as Word, in the export menu, gives you the PDF's text as a Word document without translating it.
Scanned PDFs have no text to read: run OCR on them first (for example in ABBYY FineReader), save the result as a Word document (.docx) and open that here. Pages without text in an otherwise readable PDF are left out and listed, and so are pages whose text can't be read. Running heads that change from page to page, such as chapter titles, are left out too: they aren't in the grid or the Word file. PDFs that only restrict copying or editing open normally, but a PDF that needs a password to open is refused: save a copy without the password first. Images, fonts, colours, bold and italic are not carried into the Word file, and text inside images is not translated.
Yes. In a target, press and release Ctrl (Control on a Mac too) and a list opens under the caret with suggestions for that segment: target terms from your termbases, terms marked non-translatable, the numbers, links, emails, phone numbers, codes and dates of the source (numbers also in the target language's format, so 1,500.25 becomes 1.500,25 in Spanish), fragments from your translation memories and words you've already used in this project. Type to narrow the list, use ↑ and ↓ to choose, Enter or Tab to insert and Esc to close. After 3 letters the list also opens on its own when a term, number or TM fragment from the source starts with what you typed. You can turn that, or all suggestions, off in Editor settings. The suggestions are worked out in your browser from the files you opened: nothing is sent to TraduQ or anywhere else.
Yes. Choose several files, choose a folder, or drag a folder onto the page, and they open as one project: one grid, one translation memory and one QA run, with each file in the Files panel. Bilingual files, Word, Excel, PowerPoint and PDF files, and text and software files can be mixed, up to 300 files and 200 MB in all. A project has one language pair, which you pick once: if the bilingual files have different pairs, you choose one, and the bilingual files with another pair are left out. So are files in a folder or with a name for another language than the source (such as locales/fr when you translate from English), hidden files, packages and files the editor doesn't open: the list after opening says why for each one. Download the selected file from the Files panel, or everything as one .zip with the files in their folders (a folder named after the source language, such as locales/en, comes back named after the target). The project is saved in this browser like a single file: use Resume below the drop zone, or pick the same files again, to pick up where you left off.
Yes. From the Export menu of a Word, Excel, PowerPoint, PDF or text project (or several files opened together), download it as XLIFF 1.2 (for Trados, memoQ, Phrase and most CAT tools), XLIFF 2.0 or SDLXLIFF (beta, for Trados Studio). Translate it there, then choose Import translated XLIFF and pick the file you got back: you see each change and choose what to apply, then export the translated document from TraduQ. With Include the original file on (up to 20 MB), the XLIFF carries the original document, so dropping the XLIFF alone into TraduQ offers to open the document with its translations.
Yes, when the package carries the Word original: a Trados package (.sdlppx) or SDLXLIFF that embeds its .docx, or a memoQ package (.mqout, .mqxlz) with the document inside. Export › Translated document (.docx) fills the original with your translations and, after the download, lists what stayed in the source language (untranslated, or paragraphs that didn't line up) and any formatting it couldn't place. TraduQ's segmentation can differ from Trados or memoQ, so a paragraph they split differently is filled as one segment. Open the document in Word and check it before delivery; your client may still expect the return package or the translated MQXLIFF files. The item doesn't show when the original isn't in the package (a Trados file that only points to it), or for Excel and PowerPoint originals.
Yes, in Beta, as an IDML file (.idml): in InDesign, choose File › Export › InDesign Markup (IDML). The editor translates the text of the pages, the text on parent pages and, listed last, the text off the page; hidden layers and conditional text are left out, and so is text in images. Export › Translated InDesign file (.idml) gives back the same file with only the translated text changed (named <name>_<language>.idml), and you open it in InDesign. .indd, .indt and .inx files can't be opened: export them as IDML first. The "May not fit" check, in QA and before exporting, warns about text frames whose translation is much longer than the source. It is an estimate from text length, since fonts and frame sizes aren't measured, so check the layout in InDesign after opening the file.