33 / Convert · PDF → EPUB

PDF to EPUB

Turn a text PDF into an EPUB that reflows.

A PDF is a page description, not a book: the text is placed line by line, so an e-reader cannot reflow it. This tool reads the text layer, clusters glyph runs back into lines, joins lines into paragraphs, and uses relative font size to decide what is a chapter and what is a section. Repeating lines in the top and bottom margins — running heads, confidentiality stamps, page numbers — are recognised and removed, with a count so the omission is visible. Images are pulled out as separate files in page order, and rows whose columns line up are rebuilt as tables. If the PDF has bookmarks, they win over font-size guesses and become the contents; otherwise the headings decide. The result is an EPUB 3 file with a navigable table of contents, built in this tab. The original PDF is never uploaded or changed.

All tools
Opening your workspace…

From file to finished

How to use this tool

  1. Choose a text-layer PDF from this device.

  2. Check the chapter preview and contents, then set the book title, author and language.

  3. Download the EPUB 3 file and open it in your e-reader or the EPUB reader on this site.

Lines become paragraphs, not pages

Glyphs are grouped into lines by vertical position, then lines are joined into paragraphs until a short line, an indent or a new heading ends them. End-of-line hyphenation is repaired, so “well-” and “known” become “well-known” instead of two broken words. Reading order follows the page, not the file’s internal object order.

Chapters from font size, or from bookmarks

Text at 18 pt and above becomes a chapter heading; 14–17 pt becomes a section heading inside the current chapter. When the PDF has at least two usable bookmarks, those are used instead: top-level entries become chapters and nested entries become sections, so a five-bookmark PDF gives exactly five contents entries. Dot leaders and page numbers are stripped from bookmark titles.

Running heads and page numbers are removed

A line in the top or bottom 8% of a page that repeats on other pages is treated as a running header or footer and left out of the book, and so are standalone page numbers. The summary reports how many header lines, footer lines and page numbers were removed, so the cleanup is measurable rather than silent.

Images and simple tables are carried over

Images are extracted as PNG files inside the EPUB, in page order, and skipped when they are tiny or sit entirely in the margins. A run of rows with stable, aligned column positions becomes a real HTML table with a header row when the first row is set larger. Everything else stays as paragraphs.

A little more detail

Frequently asked questions

Is my PDF uploaded?

No. The PDF is opened, measured and converted in this browser. Nothing is sent to a server, no account is needed, and the original file is not changed.

Why does the tool say my PDF looks scanned?

A scanned page is a picture of text, so there is nothing to select or reflow. If the pages have fewer than 50 characters of text on average, the file is treated as scanned and no EPUB is produced. Run OCR in your scanner software or a PDF editor first, then choose the searchable PDF.

How are chapters decided?

If the PDF has bookmarks, the top-level bookmarks become chapters and nested bookmarks become sections. Without usable bookmarks, text at 18 pt or larger becomes a chapter and 14–17 pt becomes a section. A PDF with no such headings becomes a single chapter.

What happens to headers, footers and page numbers?

Lines that sit in the top or bottom 8% of the page and repeat across pages are removed as running heads or footers, and standalone page numbers are removed too. The summary gives the exact count of header lines, footer lines and page numbers that were left out.

Are images and tables included?

Images embedded in the PDF are extracted as PNG files in page order and skipped when they are smaller than 24 px or sit entirely in the margins. Simple tables whose columns line up are rebuilt as HTML tables with a header row; complex or borderless tables may stay as paragraphs.

Will the EPUB look exactly like the PDF?

No. EPUB is reflowable, so text wraps to the reader’s screen and page size is chosen by the reader. The tool keeps the structure and the words, not the PDF page design. Open the EPUB in the reader on this site to check how it flows before publishing.

What are the limits?

One PDF up to 200 MB and 600 pages. The PDF must have a text layer; OCR, encryption removal and multi-file batching are not supported. Everything runs in this tab, so a long PDF takes longer and the tab must stay open.