You need to translate a PDF — a contract, a research paper, a supplier invoice, an onboarding document for your international team. You paste the text into Google Translate, get a reasonable translation, and then spend 45 minutes trying to recreate the original layout in a new document. Tables are broken. Headers are misaligned. Page numbers don’t match. The formatting is gone.
This is the most common complaint people have when translating PDFs, and it happens because most translation workflows treat the PDF as raw text when it’s actually a complex layout document. The text is only half the problem. Preserving the structure — columns, tables, headers, footers, images, fonts — is the other half, and it’s the half most tools ignore.
Here’s how to actually translate a PDF and keep the formatting intact, depending on your situation.
Why PDF Translation Breaks Formatting in the First Place
A PDF is not a text document. It’s closer to a printed page that’s been frozen in digital form. Each character has an exact position on the page, defined by coordinates rather than by a logical flow of paragraphs. When you extract text from a PDF for translation, you’re pulling characters out of their positions. When the translated text goes back in, it’s almost always a different length than the original — German is ~30% longer than English, Japanese can be ~20% shorter — and the layout falls apart.
Tables are especially fragile. A PDF table isn’t a real table (like in Excel or HTML) — it’s a collection of text fragments and drawn lines placed at specific coordinates. Extract the text, translate it, and the cell boundaries no longer match the content. This is why copy-paste translation workflows always produce broken tables.
Method 1: Copy, Translate, and Reformat Manually
Best for: One-off translations where layout doesn’t need to match exactly.
This is what most people try first:
- Copy text from the PDF (or use a tool like Adobe Acrobat to extract it).
- Paste into Google Translate, DeepL, or another translation service.
- Manually rebuild the document in Word, Google Docs, or another editor.
The reality: This works for simple, text-heavy documents — a one-page letter, a short memo. For anything with tables, multi-column layouts, or precise formatting, you’re looking at significant manual work. A 10-page report with tables can take 2-3 hours to reformat after translation.
When this makes sense: When you only need the translated content (not a formatted document), or when the original is simple enough that reformatting is trivial.
Method 2: Use Google Docs as a Bridge
Best for: Documents under 10 pages with moderate formatting.
Google Drive can open PDFs as editable Google Docs. The conversion isn’t perfect, but it preserves more formatting than a raw text copy:
- Upload the PDF to Google Drive.
- Right-click → Open with → Google Docs.
- Google Docs converts the PDF to an editable document (tables, headers, and basic formatting usually survive).
- Use Google Docs’ built-in translation: Tools → Translate document.
- A new translated copy is created, preserving the document structure.
The catch: Google’s PDF-to-Docs conversion struggles with complex layouts — multi-column pages, text wrapped around images, nested tables, and custom fonts often break. Simple single-column documents with basic tables translate well. Anything more complex and you’re back to manual cleanup.
Translation quality: Google Translate handles common language pairs well (English ↔ Spanish, French, German, Portuguese) but can be rough on technical or legal terminology. For formal documents, you’ll want to review the output carefully.
Method 3: Translate the PDF in Place
Best for: Preserving exact layout, especially for multi-page documents with complex formatting.
The cleanest approach is to translate the text inside the PDF without extracting it. Tools that do this replace text at its original coordinates, adjusting font size or spacing as needed to fit the translated content within the existing layout.
Super PDF Editor does this directly in Google Drive. You open the PDF from Drive, select text or entire pages, choose a target language (100+ supported), and the translation replaces the original text in place. Tables, headers, and page layout stay intact because the text is being swapped at the coordinate level, not extracted and reimported.
This matters for documents where the formatting is part of the meaning — contracts where clause numbers must align, research papers where figures and citations reference specific page positions, or compliance documents where the translated version needs to be a visual match to the original.
Adobe Acrobat Pro can also edit PDF text in place, but it doesn’t have a built-in translation feature. You’d need to translate the text elsewhere and then manually replace each text block in Acrobat — feasible but time-consuming for long documents.
Method 4: AI Translation Tools (DeepL, ChatGPT)
Best for: High-quality translations of extracted text, especially for technical or nuanced content.
AI translation has gotten remarkably good. DeepL in particular produces translations that read naturally and handle context well. ChatGPT and Claude can translate with additional instructions (“translate this legal contract maintaining formal tone” or “translate this medical report, preserving technical terminology”).
The limitation is the same as Method 1: these tools translate text, not documents. You still need to handle the formatting separately. However, for documents where translation quality matters more than layout — a patent filing, a clinical trial report, a legal brief — using an AI translator for the text and then placing it back into the PDF with a tool like Super PDF Editor gives you the best of both worlds.
| Method | Formatting preserved | Translation quality | Time for 10-page doc | Best for |
|---|---|---|---|---|
| Copy + Google Translate | None (manual rebuild) | Good for common languages | 2-3 hours | Simple one-pagers |
| Google Docs bridge | Partial | Good | 30-60 min with cleanup | Moderate formatting |
| In-place PDF editing | Full | Depends on engine used | 10-20 min | Complex layouts, tables |
| AI translation + PDF editor | Full (with editor step) | Excellent | 20-40 min | Technical/legal docs |
Common Language-Specific Issues
Not all translations are equal when it comes to formatting impact. Some language pairs cause more layout problems than others:
Languages that expand text significantly:
- English → German: ~30% longer
- English → French: ~15-20% longer
- English → Finnish: ~30-40% longer
These expansions mean translated text often overflows table cells, breaks column layouts, and pushes content onto additional pages. An in-place editor that auto-adjusts font size or spacing handles this far better than a manual rebuild.
Languages that change text direction:
- English → Arabic: Right-to-left
- English → Hebrew: Right-to-left
RTL languages require the entire page layout to be mirrored. Most PDF translation approaches can’t handle this automatically. For RTL translations, you may need to rebuild the document in a word processor that supports RTL layout, then export back to PDF.
Languages with different character sets:
- English → Chinese, Japanese, Korean: Different font requirements
- English → Hindi, Thai, Bengali: Complex script rendering
These translations require fonts that support the target script. If the PDF uses a font that doesn’t include the translated characters, you’ll see blank boxes or question marks. Make sure your translation tool embeds appropriate fonts or falls back to a universal font like Noto Sans.
What to Check After Translation
Regardless of which method you use, run through this checklist before sending the translated PDF:
- Table alignment. Are cell contents still within their borders? Do column widths make sense with the translated text?
- Page breaks. Did text expansion push content to new pages? Are headers and footers still correct on each page?
- Numbers and dates. Were they translated or left as-is? Date formats vary by locale (MM/DD/YYYY vs. DD/MM/YYYY vs. YYYY-MM-DD). Currency symbols and number formatting (commas vs. periods for decimals) may need manual adjustment.
- Proper nouns. Names, company names, and brand names should typically stay in their original language unless there’s a standard localized version.
- Legal and technical terms. Machine translation can mistranslate terms of art. If the document is a contract, medical record, or regulatory filing, have a native speaker review the terminology.
- Images with text. Diagrams, charts, and screenshots with embedded text won’t be translated by any automated tool. You’ll need to handle these manually or recreate them.
The Bottom Line
The best method for translating a PDF depends on what you need to preserve. If you just need the content in another language and formatting doesn’t matter, copy-paste into DeepL or ChatGPT and you’re done in minutes. If the layout matters — and for professional documents it almost always does — use a tool that translates text in place within the PDF, like Super PDF Editor, so you’re not spending hours rebuilding tables and margins that were already correct in the original.
The days of choosing between “good translation” and “preserved formatting” are ending. You can have both — you just need to pick the right tool for the job.