How to convert a document with Cyrillic or Greek text to PDF

Updated 19 August 2026

If a Word-to-PDF converter turns Cyrillic or Greek text into question marks, it is drawing with a font that has no glyphs for those letters. UnboundPDF’s Word, PowerPoint and HTML to PDF tools detect text outside the basic Western set and embed a real font for it — Cyrillic, Greek and extended Latin come out as text, with zero “?”. Chinese, Japanese, Korean, Arabic and emoji are not covered yet; the tool counts and names those characters rather than hiding the problem. Everything runs in your browser.

Word-to-PDF converters that print question marks for Cyrillic or Greek have no glyphs for them. How UnboundPDF embeds a real font instead.

1. Open Word to PDF (or PowerPoint to PDF / HTML to PDF for those sources) and drop in the file.
2. Convert. Text in Cyrillic, Greek or extended Latin is embedded with a real font automatically.
3. Read the result note: if any characters could not be covered, it says how many and which.
4. Download; select the text in any reader to confirm it is real text, not a picture.

Why converters print “?”

A PDF has to carry, or refer to, a font for every character it shows. The simplest converters use one of the standard built-in fonts, which cover Western European letters and not much else. Ask such a font for a Cyrillic “Д” or a Greek “λ” and there is no glyph — so the converter writes the only thing it can: a question mark, or a blank. The document looked fine in Word because Word had the font; the PDF does not.

What changed

Our three converters now check each run of text. Plain ASCII and ordinary accented Western text — café, naïve, Grüße — still go through the fast built-in path with nothing extra to download. Text outside that set triggers a real font: the tool loads an open-licensed family from our own site, subsets it to the characters you used, and embeds it. Cyrillic, Greek and extended Latin (Polish, Czech, Turkish, Romanian and the like) all come out as selectable text with zero substitutions. Bold and regular runs each get the right face.

What is still not covered

Chinese, Japanese and Korean, Arabic, and emoji. The embedded family has no glyphs for them. When such characters occur the result note counts them and lists the code points — “3 characters could not be covered: U+0639, U+6F22, U+1F600” — instead of printing “?” or dropping them quietly. Work on CJK coverage is planned but not shipped; treat those scripts as unsupported today.

Layout is unchanged

The font change does not alter line breaks, page counts or positions for documents that were already converting correctly — measured on our corpus as byte-identical output before and after. You get the same page, with the letters you typed.

Which converter for which file

Word to PDF for .docx. PowerPoint to PDF for .pptx. HTML to PDF for web pages and pasted HTML. All three embed fonts the same way and none uploads your file. If a PDF you already have shows missing or wrong letters, PDF fonts wrong or missing explains what happened.

Frequently asked questions

Why does my Word-to-PDF conversion show question marks instead of Cyrillic?

The converter drew the text with a font that has no Cyrillic glyphs. UnboundPDF detects such text and embeds a real font, so Cyrillic, Greek and extended Latin come out as text with no “?”.

Does Chinese or Japanese text convert correctly?

Not yet. CJK, Arabic and emoji are not covered; the tool counts and names the characters it could not render instead of hiding them.

Does embedding a font change my layout?

No. For documents that already converted correctly, output is byte-identical before and after. Only the previously missing letters change.

Is the document uploaded to convert it?

No. Conversion and font embedding run in your browser; the font files are fetched from our own site only when a document needs them.