PDF is the format most people have, and the hardest one to translate. A PDF is built to look the same on every screen. It stores where each word sits on the page, not which words form a paragraph. Before a PDF book can be translated, the text has to be read back out of the page. This guide explains which PDFs work, what happens to the layout, and when another file is the better choice.
First check: does your PDF have a text layer?
There are two kinds of PDF.
- Digital PDFs come from a word processor or a publishing tool. You can select and copy a sentence. These can be translated.
- Scanned PDFs are photos of printed pages. Nothing on the page can be selected, because there is no text, only an image. These need OCR (optical character recognition) first, and OCR makes errors that a translation then repeats.
A quick test: open the PDF, try to select one sentence and paste it into a text editor. If the sentence comes out clean, the PDF has a text layer. If nothing is selected, or the pasted text is garbled, the PDF is scanned or damaged.
We accept only PDFs with a text layer. A scanned PDF is refused at upload, since we do not run OCR.
What happens to the layout
A translated sentence rarely has the same length as the original. Spanish and French often run longer than English, for example. So a translated PDF cannot keep every line in the same place. The text has to flow again.
Our service sets the translated text again on pages of the same size, with the same margins, and with the same font or a close one. Headings, paragraphs and images stay in their order. The page count can change, because the translated text is longer or shorter.
Some things in a PDF are hard to recover in any tool:
- Running headers, footers and page numbers are mixed into the text on every page. They have to be removed, or they end up in the middle of a sentence.
- Columns and sidebars can be read in the wrong order.
- Words split at the end of a line with a hyphen have to be joined again.
- Tables, formulas and text inside pictures usually need a manual check, and text inside a picture is not translated.
A novel with plain pages comes through well. A textbook with columns, boxes and tables needs more checking.
Use EPUB or DOCX when you can
If you have the book in another format, use it. An EPUB or a DOCX file stores real paragraphs, chapters and headings, so nothing has to be guessed.
- EPUB in, EPUB out: the original chapters and table of contents are kept.
- DOCX in, DOCX out: headings and styles are kept.
- PDF in, PDF out: the text is set again, as described above.
If you wrote the book yourself, export it from your word processor as DOCX instead of PDF.
A checklist for a PDF book
- Confirm that you may translate the book (public domain, your own work, or written permission).
- Select a sentence to check that the PDF has a text layer.
- Look for an EPUB or DOCX version of the same book first.
- Run the free sample and compare the first pages with the original.
- After the full translation, check chapter starts, tables and image captions.
Ready to try? Upload your book and read the first 5 pages for free.
Related reading
- See what the engine handles for each language pair: English to German, German to English and English to Portuguese.
- Read the translated file on your device: reading a translated book on Kindle, Kobo and other e-readers.