Does OpenImmersive translate scanned PDFs?

Not yet. Scanned, image-only PDFs with no text layer aren't supported — OCR support is on the roadmap. If you can already select text in the original PDF, it will work.

Is the translated text a flattened image, or can I still select and search it?

It's real, selectable, searchable vector text — not a flattened image. You can copy, search, and highlight the translated text the same way you would the original.

What happens to tables and numbers when a PDF is translated?

Tables are detected automatically. Numbers and figures keep their native alignment; only the text labels are translated in place, so the table structure isn't broken.

Does my PDF get uploaded anywhere?

No. Everything runs locally in your browser. The text you select is sent directly to the translation engine you chose — nothing is uploaded to OpenImmersive's servers, because there are none.

What's the difference between bilingual and translation-only mode?

Bilingual mode alternates original and translated pages side by side, for comparison. Translation-only mode opaquely covers the original text at its original background color, so only the translation is visible, while image layout stays intact.

PDF Translation That Keeps Your Layout Intact

Most “PDF translation” tools do one of two things badly: they either flatten your document into a translated screenshot, or they translate the text and leave tables, images, and layout scrambled. OpenImmersive does neither — it keeps the PDF’s native structure while replacing only the text.

Key facts

What	Detail
Output text	Real, selectable, searchable vector text — not a flattened image
Images & graphics	Native PDF pages preserved — images, vector graphics, table borders and shading stay 100% intact
Tables	Detected automatically — numbers and figures keep their native alignment, only text labels are translated in place
Where translation happens	Locally in your browser — nothing is uploaded
Known limitation	Scanned, image-only PDFs (no text layer) aren’t supported yet

Last updated: 2026-07-04

Two modes, depending on how you read

Bilingual — original and translated pages alternate, side by side, so you can compare sentence by sentence. This is the mode most people reach for when reading a paper or contract they need to get exactly right.

Translation-only — the translated text opaquely covers the original at its original background color, so only the translation is visible. The source text isn’t visible, but image layout is preserved underneath. Useful when you just want to read the document in your language without the visual noise of two versions.

Why layout usually breaks (and why this doesn’t)

Most PDF translators work by extracting all the text, running it through a translation engine, and then re-flowing it into the page — which is exactly what breaks tables, multi-column layouts, and image captions. OpenImmersive instead translates text in place: it finds the existing text objects in the PDF, swaps their content, and leaves every other object (images, vector graphics, table borders, shading) untouched. Numbers and figures in tables are specifically left alone — only the text labels are localized — so a table of financial figures or scientific data doesn’t shift when the units and labels change length between languages.

Where it doesn’t work yet

If a PDF is a scanned image with no underlying text layer, there’s no text to translate — the extension can’t invent an OCR pass yet, so scanned-only documents aren’t supported. Text embedded inside images (screenshots of text, diagrams with labels baked into the image) is skipped for the same reason. Both are on the roadmap.

Key takeaways

Translated PDFs stay selectable, searchable, and copy-pasteable — they’re text, not pictures.
Tables keep their shape: only labels translate, numbers and alignment don’t move.
Nothing about your PDF is uploaded to a server OpenImmersive controls, because there is no server.
Scanned/image-only PDFs are the one gap today.