Translate PDF documents
Translate any PDF into 100+ languages while preserving layout, fonts, and images. Includes Spanish, French, Arabic, Mandarin, and more.
Translate PDF is coming soon
We're putting the finishing touches on this one. Sign up and we'll email you when it's ready.
Translating a PDF traditionally meant copy-pasting text into Google Translate, which strips all formatting and gives you an unstructured wall of translated text. PDFOnly's PDF translator preserves the original layout — headings stay headings, tables stay tables, images stay in place — and substitutes the translated text into the existing document structure. Output is a polished, ready-to-share translated PDF that looks like the original but reads in the target language. Supports 100+ languages including Spanish, French, German, Arabic, Mandarin, Japanese, and Russian.
How to translate pdf step by step
- 1
Upload your PDF
Drop any PDF up to 100 MB free, 200 MB on Pro. Digital and scanned PDFs both work.
- 2
Pick source and target languages
Source can be auto-detected (recommended) or specified. Target is your output language. For mixed-language source documents, specify the primary source.
- 3
Choose output style
Translated-only (target language replaces source) or bilingual (target alongside source on each page). Bilingual is useful for review or learning.
- 4
Click Translate and download
Translation runs page by page, preserving the layout. Output is a fresh PDF ready to share.
Why translate pdf on PDFOnly
Layout preservation
Most online PDF translators strip formatting and dump the result as plain text. We substitute translations back into the original document structure so your output looks like a real translated document, not a stripped transcript.
Frontier translation models
We use GPT-class and Claude-class translation, which significantly outperforms older statistical machine translation on idioms, technical terminology, and long-context coherence.
Privacy-first
Documents are never used to train translation models. Confidential contracts, medical records, and internal documents stay yours.
What people use translate pdf for
A few common scenarios. If your workflow looks like one of these, this tool is a good fit.
Translate a vendor contract for legal review
Foreign-language contract becomes English (or vice versa) with formatting intact, so your lawyer can review the actual document — not a stripped-down text dump.
Localize technical documentation
Convert product manuals or API docs into your customers' languages while keeping diagrams, code samples, and tables in place.
Read foreign research papers
Researchers reading literature in a non-native language get a faithful translated PDF they can annotate and cite alongside the original.
Translate scanned historical documents
Old letters, archival records, or family documents in another language become readable. OCR runs first on the scan, then translation.
What you get
- 100+ supported languages with high-quality neural translation
- Original layout preserved — headings, tables, images, page breaks intact
- Bilingual output option — original and translation side by side
- Handles both digital and scanned PDFs (OCR runs first on scans)
- Domain-specific glossaries on Team plan (legal, medical, technical terms)
- Files auto-deleted within an hour, never used to train AI models
Frequently asked questions
How accurate is the translation?
For major language pairs (English ↔ Spanish/French/German/Mandarin/Japanese): excellent for general prose, very good for business and technical content, occasionally needs human review for legal nuance. For low-resource pairs (e.g. Maltese to Tagalog): less reliable, post-edit recommended.
Will the translated PDF look like the original?
Layout, headings, tables, and image positions are preserved. Text length differs between languages (German is often longer than English; Chinese is often shorter), so individual paragraphs may grow or shrink — we re-flow text within the original page layout.
What languages do you support?
100+: all major European languages, Arabic, Hebrew, Hindi, Bengali, Tamil, Mandarin (Simplified and Traditional), Japanese, Korean, Vietnamese, Thai, Indonesian, Malay, Swahili, and many more. Full list in the language picker.
Can I translate scanned PDFs?
Yes. We OCR scans automatically before translating. OCR accuracy directly affects translation quality — a clean 300 DPI scan translates much better than a faded photocopy.
Will the translation be 100% identical to a human translator?
No — even the best AI translation has stylistic and occasional accuracy gaps a professional translator catches. For high-stakes documents (contracts, regulatory filings, medical records), have a native speaker review the output. For internal use, research, or rough localization, AI is more than sufficient.
Can I provide custom terminology?
Team plan supports domain glossaries — upload a CSV of source-target term pairs (e.g. company-specific product names that shouldn't be translated, or specific legal terms). The translator honors them across all your translations.
Will my document be used to train translation models?
No. We use API tiers from frontier model providers with explicit no-training-data clauses. Documents are sent for inference and discarded.
Ready to translate pdf?
Free to use for the basics. Files are auto-deleted within an hour and never used to train AI.
Open Translate PDF