Document to PDF

Reproducing a Word document exactly means reproducing Word: the same font metrics, the same line-breaking rules, the same table engine, the same pagination. Services that manage it run a copy of LibreOffice on a server — which is why they need you to upload the file.

Nothing runs on a server here, so the honest ceiling is a clean re-typesetting. The words, the structure and the emphasis are yours; the line breaks and page breaks are ours. For most documents — a report, notes, a letter, a CV — that is fine and often tidier. For anything where the layout is the content, print to PDF from Word instead, which is exact and takes about the same number of clicks.

Text is set in the standard PDF fonts, which every reader is guaranteed to have and which cover WinAnsi — English, western European accented letters, curly quotes, dashes, the euro sign. Greek, Cyrillic, Chinese, Japanese, Korean, Arabic, Hebrew and emoji are outside it.

If your document contains any, they are counted and named before the file is written rather than causing a failure partway through, and you can decide whether to continue. Embedding a font that covers them would make every PDF several hundred kilobytes heavier, which is the wrong trade for the great majority of documents.

.doc — the old binary Word format, replaced in 2007. Reading it needs a parser for a specification Microsoft never fully published, and nothing in a browser has one. Open it in Word, LibreOffice or Google Docs and save as .docx.

.odt, .rtf, .pages — not supported. All three export to .docx from the applications that make them.

Password-protected .docx — refused rather than guessed at.