PDF to EPUB: Convert a PDF Into a Reflowable Ebook

November 11, 2026 · 10 min read

PDF to EPUB: How to Convert a PDF Into a Reflowable Ebook

PDF to EPUB conversion turns fixed pages into text that reflows to fit any screen. For a clean, text-heavy PDF, run it through Calibre with tuned PDF Input settings. For a messy one, extract the text to Markdown, fix the chapters, then build the EPUB yourself. Either way, test the result on the device you read on.

Why PDF to EPUB Conversion Loses Formatting

A PDF places every glyph at a fixed position on a fixed page. It doesn't record which lines form a paragraph, which text is a heading, or which of two columns you read first. An EPUB works the other way: the W3C's EPUB 3.3 Recommendation says publications are meant to reflow to the available screen by default.

So every converter has to guess. It measures line lengths, font sizes and gaps, then rebuilds paragraphs and headings from those clues. Calibre's conversion manual is blunt about it. It calls PDF one of the worst formats to convert from and lists what fails:

  • Multi-column and image-based documents aren't supported.
  • Tables and vector images aren't extracted.
  • Links and the PDF's own table of contents don't carry over.
  • Ligature glyphs for fi, ff or ll may or may not convert.
  • Right-to-left scripts and math typesetting won't convert correctly.

That's the honest answer if you hope to convert PDF to EPUB without losing format. The page layout goes, by design. What you can keep is the text, the headings and the reading order, as long as the source is clean.

Which PDFs Convert Well and Which Should Stay PDF

Before you convert anything, decide whether the file should reflow at all. Ask what a reader loses when the printed page disappears.

PDF typeMake it a reflowable EPUB?
Novel, essay or report with a real text layerYes, the ideal case
Scanned book where each page is a photoOnly after OCR adds a text layer
Textbook with sidebars, figures and two columnsUsually not
Academic paper full of equationsUsually not
Comics, sheet music, forms, slide decksNo

Here's a quick test. Try to select a sentence in your PDF viewer. If you can't, the page is probably an image, with no text layer for a converter to read.

We prefer leaving layout-heavy PDFs as PDFs. A tablet displays them well enough, and a forced conversion often reads worse than the original. EPUB does have a fixed-layout mode for pre-paginated content, but that's a different job from reflowing text.

How Do I Transfer a PDF to an EPUB With Calibre?

Calibre is a free, open-source ebook manager for Windows, macOS and Linux. A Calibre PDF to EPUB conversion runs entirely on your own machine, so nothing gets uploaded. In the app:

  1. Click Add books and pick your PDF.
  2. Select the book in your library and click Convert books.
  3. Set the output format at the top of the dialog to EPUB.
  4. Open PDF Input and find the line un-wrapping factor. The default is 0.45. Lower it to join more lines into paragraphs, or raise it to join fewer.
  5. Open Search & replace and add a regular expression that matches your running header or footer, so it doesn't land mid-paragraph.
  6. Click OK, then open the finished book with View and read a few chapters.

The same conversion runs from the command line with ebook-convert, which comes with Calibre:

ebook-convert report.pdf report.epub --unwrap-factor 0.3

That writes report.epub and joins more short lines into paragraphs than the default setting would. Expect a few rounds of adjusting. Calibre's manual warns that PDF output ranges from decent to unusable, depending on the input file.

Rebuilding a Messy PDF as a Clean EPUB Through Markdown

When Calibre's guess isn't good enough, take over the structure yourself. Markdown makes a good middle step because it's plain text, its headings are explicit, and an EPUB builder can split chapters on them. On our site this is two separate tools, not a one-click PDF to EPUB converter.

  1. Extract the text. Upload the PDF to PDF to Markdown. It runs on our server and reads only the text layer, so scanned pages give little or nothing and embedded images aren't extracted. On the free plan, each conversion uses one of 3 server operations per day. Our guide to converting PDF to Markdown compares other extractors and covers OCR for scans.
  2. Fix the structure. Paste the result into the editor and clean it up while the preview updates on the right.
  3. Build the book. Feed the cleaned Markdown to an EPUB builder, covered two sections down.

You still lose the page layout on this route. What you gain is a book with real chapters, whole paragraphs and no stray page numbers, because you decided where each one goes.

Cleaning Up Extracted Text Before It Becomes Chapters

Extracted text drags the PDF's page furniture along with it. Four fixes handle most books.

Delete running headers and page numbers. They repeat on every page and cut paragraphs in half.

THE HARBOUR                                  47

The boats came in before dawn, low in the water
and loud with gulls.

48

Remove the header line and the bare 47 and 48. Then join the paragraph's two wrapped lines into one.

Rejoin words split across lines. Print layout hyphenates words at the right margin.

The harbour master kept a care-
ful record of every boat.

That becomes one line: The harbour master kept a careful record of every boat. Watch for genuine hyphens, like well-known, that happen to fall at a line end.

Turn chapter labels into headings. Extracted text often treats CHAPTER 3 as a short paragraph. Make it a level-1 heading so the EPUB builder starts a new chapter there.

# Chapter 3: The Storm

Delete the printed table of contents. Its page numbers mean nothing once text reflows. EPUB readers build their chapter list from the book's navigation instead.

Here's the shape you're aiming for: each H1 starts a chapter, each H2 is a section inside it, and every paragraph sits on its own unbroken line. Our editor shows a single line break inside a paragraph as a visible break, so leftover PDF line breaks are easy to spot in the preview.

Chapter 1: The Harbour

The boats came in before dawn, low in the water and loud with gulls. Nobody on the quay looked surprised.

Mara counted them twice. Eleven, not twelve.

A Note on the Tides

The tide table pinned by the door listed three things:

  • high water at 5:40
  • low water at 11:52
  • the date of the next spring tide

The next chapter would start with another H1 line.

75 words389 characters15 lines
Markdown

Building the EPUB From Markdown

You have two good ways to package the cleaned text.

In the browser. Paste it into Markdown to EPUB. It runs entirely in your browser, so there's no daily limit. It splits chapters at H1 (or H2), lets you add book metadata and upload a cover image, and writes an EPUB 3 file. One catch: images referenced by URL in your Markdown aren't fetched or embedded. Only the cover goes inside the book.

With Pandoc. For scripted or repeatable builds, Pandoc's epub output is EPUB 3:

pandoc book.md -o book.epub --metadata title="The Harbour" --epub-cover-image=cover.jpg

That builds book.epub with a title and cover. Pandoc splits the internal files at level-1 headings by default, and --split-level changes that. It also pulls images your Markdown references into the EPUB. Our Pandoc Markdown post covers its syntax extensions and other commands.

Can You Read a Converted EPUB on Kindle?

Yes. Amazon's Send to Kindle help page lists EPUB as a supported file type, alongside PDF, TXT, HTML, RTF and common image formats. As of November 2026, MOBI isn't on that list. You can send the EPUB through Send to Kindle for Web, by email, or from the Kindle app. It then lands in your Kindle library.

Our Markdown to EPUB tool has no Kindle or MOBI output, and for this workflow you don't need one. You can also send the original PDF as is. By default it stays a PDF, though, with fixed pages that don't reflow, so on a small e-ink screen you'll zoom and pan. That's the usual reason a PDF to EPUB Kindle workflow is worth the effort.

Checking the EPUB Before You Share or Read It

PDF conversions fail in quiet ways. A chapter can vanish into the previous one, or a page break can leave half a sentence stranded. Compare the EPUB against the original PDF, not just against your memory of it:

  • Put the chapter list next to the PDF's printed contents page. Every chapter should appear once, in the same order.
  • Find a few sentences that run across a page break in the PDF and search for them in the reader. Each should read as one unbroken sentence, with no stray number or header inside it.
  • Shrink the reader to phone width and raise the font size. Text should rewrap cleanly with no fixed-width lines left over.
  • Validate the package with the W3C's EPUBCheck, which needs Java installed:
java -jar epubcheck.jar book.epub

The validator prints each error and warning with its location in the file. Clear the errors before you hand the book to anyone else.

One limitation applies to every route here. Neither Calibre nor the Markdown route recovers tables or illustrations reliably from a PDF. Calibre doesn't extract tables or vector images, and our PDF to Markdown tool skips images entirely. For an illustrated book, plan to add figures back by hand, or keep the PDF.

Free PDF to EPUB Converters and Other Questions

A good PDF to EPUB result starts with a decision, not a converter. Keep layout-heavy files as PDFs, run clean text through Calibre with a tuned unwrap factor, and rebuild messy ones through Markdown so every chapter break is yours. When your Markdown is ready, paste it into Markdown to EPUB, add a cover and metadata, then open the finished book on your reader before you call it done.