DocuHonyaku

DocuHonyaku is a document translation service built around a single promise: the translated file looks like the original, just in a different language. The workflow is upload-and-download — drag a file onto the interface, pick a target language, and receive the translated document in the same format. No manual reformatting, no copy-pasting into a new template. The vendor states files up to 1 GB are accepted, and supported formats span PDF, DOCX, XLSX, PPTX, scanned PDFs, image files, EPUB, and SRT/VTT subtitle files.

The differentiating claim is layout preservation at a structural level rather than a cosmetic one. The vendor describes a four-stage pipeline: structural analysis to map columns, tables, figures, and equations; AI-based OCR (not traditional character recognition) for scans so context informs reading; LLM-based translation that carries meaning across sentence boundaries; and a reconstruction step that writes the output back into the original file format. A before/after example on the vendor page shows a two-column English academic paper — equations in LaTeX-style notation, figure captions, footnotes — rendered in Japanese with the same spatial arrangement.

This fits research and business workflows where a document goes directly back into use after translation: a grad student reading a paper alongside the original, a manufacturing team distributing a localized Excel spec sheet, a legal office reviewing a scanned contract draft. It does not fit teams that need to automate translation through an API, integrate with a document management system, or keep files entirely within their own infrastructure — there is no API, no integration layer, and no self-hosted option. The 72-hour auto-deletion policy and stated no-training commitment address compliance questions for many teams, but regulated industries requiring on-premises data handling have no path forward here.