doc-html-translate

About

Translate PDFs and EPUBs with your browser's own free Translate page.

Browsers can translate any web page for free, but not PDFs or EPUBs: the PDF viewer draws text on a canvas that translation can't read, and EPUBs don't open in the browser at all. doc-html-translate fixes both. When you open a PDF, it extracts the text with PDF.js (bundled, offline) and lays it out as clean HTML paragraphs and headings. When you open an EPUB, it unpacks the book and combines its chapters into one clean HTML document. Either way you get a table of contents from the document's own outline - and your browser's built-in "Translate page" then works exactly as it does on any website.

Why reflow instead of an overlay? A canvas/text-layer overlay leaves the original glyphs under the text, which native translation doubles and garbles. Clean reflow gives the browser a normal DOM it can translate perfectly.

Features
- Works on web (https) and local (file://) PDFs and EPUBs.
- Detects the document language and sets it so the browser offers the right translation.
- Reading controls: font size and family, light / sepia / dark themes, jump to page, collapsible table of contents.
- One click back to the original PDF in the native viewer.
- Turn it off globally or per-site from the toolbar.

Privacy
Everything runs on your device. The extension has no servers, no telemetry, no accounts, and loads no remote code. It never uploads your documents. The only network access is your browser's own translation feature, which you trigger yourself.

Limitations
- Scanned / image-only PDFs have no text to translate; the extension detects this and offers the original (run OCR first, e.g. with the doc-html-translate desktop app).
- Files served without a ".pdf" / ".epub" address aren't auto-detected - open them from the toolbar's "Open file".
- EPUBs are shown with the viewer's own clean reading style rather than the book's original design.

Open source: github.com/SerZhyAle/doc-html-translate