Translate Word into 100+ languages — tables, lists and footnotes stay in place
Drop a .docx, pick a target language, and download a Word document that looks like the original — only the words have changed. We rewrite at the OOXML level, so tables, multi-level numbering, hyperlinks, headers, footers, footnotes, comments and track changes all survive intact.
Drag & drop a PPT, Word (.docx), or PDF here, or click to browse
Free trial · No signup · No credit cardMax 100 MB · Files are deleted within 24 hoursUsed daily for contracts, reports, academic papers, manuals and product specs.
Why translating Word is uniquely difficult
Tables and merged cells get broken
Most tools extract text into a flat stream, losing column widths, merged cells and nested tables. The translated document needs to be re-laid-out by hand.
Multi-level numbering and references break
Numbered lists, cross-references and field codes (page numbers, TOC) are easy to corrupt. The translated file shows wrong numbers or empty links.
Headers, footnotes and comments are dropped
Page headers, footers, footnotes, endnotes and review comments live in separate parts of the .docx package. Generic translators ignore them, so they come back in the source language — or vanish entirely.
Built specifically for Word
OOXML-level rewriting
We rewrite text at the run level inside the original document.xml, preserving every w:rPr style (font, color, bold, italic). Tables, merged cells and nested tables stay intact.
Hyperlinks, lists, fields preserved
Hyperlink display text is translated while the destination URL is left untouched. Multi-level numbering, cross-references and field codes are left to Word to render so numbers stay correct.
Headers, footers, notes, comments translated
Every header, footer, footnote, endnote and comment part is walked and translated — nothing comes back in the source language.
Track changes survive
Inserted text (w:ins) is translated normally; deleted text (w:del) is preserved as-is so the revision history stays intact. Math (m:oMath) is preserved byte-for-byte.
TransKeep vs. other Word translation methods
| TransKeep | Generic Translators | Manual Translation | |
|---|---|---|---|
| Tables & merged cells | Preserved in place | Torn apart | Manually rebuilt |
| Multi-level numbering | Auto-rendered by Word | Corrupted | Manual renumbering |
| Headers / footers / footnotes | Fully translated | Dropped or in source language | Manual copy |
| Hyperlinks | Display translated, URL preserved | Lost | Manual re-link |
| Track changes & comments | Preserved + translated | Stripped | Manual review |
| Output format | Native editable .docx | PDF dump | Manually re-exported |
Who uses TransKeep for Word
Legal & compliance
Translate contracts, NDAs and policies while keeping numbered clauses, definitions and signature blocks intact.
Research & academia
Translate theses, journal articles and grant proposals with citations, footnotes and equation formatting preserved.
Product & technical writing
Localize manuals, SOPs and specification documents without redoing tables, diagrams or step numbering.
Consulting & proposals
Translate proposals and reports — tables of contents, page headers and review comments all survive.
Word translation FAQ
Modern .docx (Office 2007+). Legacy .doc (Office 97-2003 binary format) is not yet supported — convert to .docx in Word first.
Yes. We rewrite at the run level inside the original tables and let Word's auto-numbering render numbered lists, so column widths, merged cells, nested tables and numbering all stay intact.
All translated. Each header/footer/footnote/endnote/comment part of the .docx is walked and translated, not just the body.
We preserve the revision marks. Inserted text (w:ins) is translated normally; deleted text (w:del) is left as-is so the revision history stays visible when the file is reopened in Word.
Math (m:oMath) is preserved byte-for-byte so the equations stay correct. Embedded Excel/PowerPoint objects are kept but their internal content is not translated in this release.
All transfers are SSL/TLS encrypted. Files are auto-deleted within 24 hours. We never use your content to train AI models.