Translate a PDF
Get the same PDF back in another language: every picture, background, table and page exactly where it was, and each paragraph translated in its own place, in its own size, colour and weight. Nothing is stored, ever.
- Never stored
- No queue, no waiting
- No signup, no watermark
How it works
Add your PDF
Drop a PDF onto the page. Its language is recognised for you.
Choose the language
Pick the language to translate into. The first time you use a language, its language data downloads once and is kept for next time.
Translate and download
You get a PDF with the same pages, pictures and layout, with every paragraph, heading and table cell in the new language.
Translating a format that was never meant to be edited
A Word document is a stream of paragraphs; a PDF is a set of instructions for painting letters at exact positions. There are no paragraphs in it — only lines, and often only fragments of lines — so the first job is finding them again: which fragments form a line, which lines form a paragraph, which pieces of text are cells of a table. Each paragraph is then translated whole, because translating line by line breaks sentences that run across lines.
The original words are removed from the page, not covered with white boxes, so whatever is underneath — a photograph, a coloured band, a table border — shows through exactly as before, and the old text cannot be found by searching or copying. Every page is then checked: if any original words survived removal, that page is put back exactly as it was rather than shipped with two texts on top of each other.
The translation is set where the paragraph was, at its size, colour and weight. PDFs embed only the letters their fonts were first used for, so a translation is set in fonts made for its script, including joined Arabic, Bengali and Devanagari with correctly placed vowel signs, and Chinese, Japanese and Korean. Everything that is not text — images, drawings, links, page sizes, bookmarks — is the original file's.
When you want something else
If you need to rework the translated document rather than read it, convert the PDF to Word with PDF to Word and translate the Word file with Translate DOCX: the text then reflows naturally, which a PDF cannot do.
For a PDF that will be printed or published, professional desktop publishing is still the right tool — a translator and a typesetter working from the original source files will always beat reconstructing a page from its painted output.
Frequently asked questions
Is the layout really kept?
Yes. Pictures, drawings, backgrounds, colours, table borders, links and page sizes are the original file's, untouched. Only the words change: each paragraph, heading and table cell is translated as a whole and set back in its own place, in its own size, colour, weight and alignment — justified stays justified, centred stays centred.
What happens when the translation is longer?
A PDF page cannot push the text below it further down, so a longer translation first uses the empty space in its own column, and only then is set a little smaller to fit. The page tells you how many passages that happened to. German and Russian typically run 20–30% longer than English; Chinese and Japanese run shorter.
Is my PDF uploaded or stored?
No. The PDF is translated here and never sent anywhere — which matters for contracts, medical records and anything else you would not paste into a website. What downloads is language data, the same for everyone, once per language.
Does it work for Arabic, Hebrew, Bengali, Hindi, Chinese and Japanese?
Yes, in both directions. Text is set in fonts made for each script, with letters joined and vowel signs placed correctly, and paragraphs become right-aligned when translating into Arabic, Hebrew, Persian or Urdu (and left-aligned when translating out of them).
Can it translate a scanned PDF?
Not directly: a scan is a picture of a page, with no text in it to translate. Run it through [OCR PDF](ocr-pdf) first to make it searchable, then translate the searchable copy.
Why does the text look slightly different?
A PDF carries only the letters its fonts were first used for, so a translation needs letters the original fonts do not have. The text is set in a matching serif, sans-serif or monospaced face instead, at the original size, colour and weight.
How good is the translation?
It is the machine translation Mozilla built into Firefox — good for understanding and working from a document, strongest between English and the major European languages. Names, places and specialist terms can come out wrong, so have a fluent speaker check anything legal, medical or going to print.
Good to know: Scanned PDFs need OCR first, and password-protected PDFs need their password removed. Text inside pictures, filled-in form fields and comments is not translated. The original typefaces are replaced by matching standard ones, and a translation that runs longer than the original is set slightly smaller to fit its space. Machine translation is not a certified translation.
Put this tool on your website
Free for any blog, class page or help article. Paste one snippet and your visitors can use it right on your page.
Related tools
Translate DOCX
Word documents in another language, layout intact
Translate Excel
Workbooks in another language, formulas intact
PDF to Word
Convert to editable .docx, with OCR for scans
OCR PDF
Make a scan searchable
Edit PDF
Rewrite existing text, add, draw and sign
Unlock PDF
Remove a password you already know