Publicado
How to Turn a Fixed PDF Into a Reflowable EPUB
PDF locks text to a page; EPUB lets it reflow. Here's the actual process for turning a fixed PDF into a reflowable EPUB — text layer, OCR, and the cleanup that follows.
A PDF is a picture of a page. An EPUB is text that can rearrange itself. That single difference is why dragging a PDF into an EPUB reader doesn't give you a book that resizes to your screen — and why "converting" one to the other means rebuilding the content, not just renaming the file.
Quick answer
There's no lossless, one-click way to do this, because the two formats store content differently at the core. PDF places text and images at fixed coordinates on a page, the same way every time, on every device. EPUB's default layout is reflowable instead — built from XHTML and CSS, so a reading app can rewrap text to fit any screen or font size. Converting means pulling the real text out of the PDF and rebuilding it that way. Which is also why the process looks different depending on whether your PDF already has readable text underneath it. For more on why some files resize and others don't, see Reflowable vs. Fixed-Layout.
Step-by-step
Check whether your PDF already has a text layer
Open the PDF and try to select and copy a line of text. If it highlights and pastes as text, there's a text layer underneath — the PDF was exported from a word processor, a website, or a layout tool, and the actual characters are stored, not just their picture. Nothing selects? Or you're copying a smudged screenshot of words? Then you're looking at a scan: an image of a page with no text behind it yet.
Extract text directly, or run OCR first
A PDF with a working text layer can go straight to extraction — the converter reads the stored characters and rebuilds them as EPUB's XHTML. Scanned pages need optical character recognition first, which is exactly the gap what OCR is, and when you need it covers in more depth. The Internet Archive's own pipeline shows how this decision gets made in practice: it uses the PDF's existing text layer when one is reliable, and falls back to running Tesseract OCR against the page image when that layer is missing or broken.
Convert with a tool built for it
Calibre is the free, cross-platform tool most people reach for. Point it at the PDF and it either extracts the existing text layer or runs OCR, then rebuilds the result as EPUB's normal document structure — see a plain guide to EPUB 3 for what that structure contains. One thing worth knowing before you start: simply repackaging the PDF's pages as images inside an .epub file produces a fixed-layout EPUB, not a reflowable one. It'll open in an EPUB reader fine. It just won't resize — you've moved the same fixed page into a different container, nothing more.
Common problems and fixes
The output is rarely perfect on the first pass. Garbled or misspelled words usually mean the OCR misread the scan — try re-running at a higher resolution or with a different engine. Headings that turn into plain paragraphs happen when the converter had no structural markup to preserve; you may need to reapply heading styles by hand afterward. Footnotes and tables tend to scramble on reflow, since they depended on a fixed position in the original — expect to simplify or reformat them manually. Images can land in odd places too, once the surrounding text starts moving. None of this is unique to PDFs, either — converting a Word document into something you'll actually read runs into the same category of cleanup.
Doing this in Margina
Margina reads PDF, EPUB, and Word files natively, so converting isn't something you need just to get a document onto your phone. It's worth doing when you specifically want true reflow — a small screen, a larger font — and the source you have is a scan or a fixed layout you don't otherwise control. Once it's converted, the EPUB opens in Margina like anything else: no account, no upload, read straight from local storage.
Frequently Asked Questions
Will the converted EPUB look exactly like the PDF?
No — that's the point of converting. A reflowable EPUB drops the PDF's fixed page image in favor of text that can rewrap, resize, and re-font itself. Page numbers, exact line breaks, and pixel-precise spacing don't carry over. What does carry over is the actual words, headings, and images, laid out fresh for whatever screen you're reading on.
My PDF is a scan — can I still convert it?
Yes, but it needs an OCR pass first. A scanned PDF is a photo of a page; there's no underlying text to reflow until OCR turns those pixels into machine-readable characters. Tools like Calibre run OCR as part of the conversion — the more accurate the OCR, the cleaner the resulting EPUB reads.
Is this legal for any PDF I have?
It's fine for PDFs you own or that are in the public domain — a contract you signed, a report you were sent, a book past copyright. It is not a workaround for DRM-protected files or copyrighted material you don't have rights to; this only covers converting files that are already yours to convert.
Conclusion
Fixed and reflowable aren't a setting you toggle. They're two different ways of storing a document — one as a page image, one as text that can rearrange itself. Converting well means finding that text first, whether it's already sitting in the PDF's text layer or has to be pulled out with OCR, then rebuilding it as an EPUB. Once that distinction is clear, it applies beyond PDFs — most "why won't this resize" questions come back to the same fixed-versus-reflowable line.