Why conversions usually go wrong
A PDF doesn't know what a paragraph or a chapter is. It only stores where each piece of text sits on a fixed page. A converter has to rebuild the book's structure from those positions, and simple converters get it wrong in predictable ways:
- Every line becomes its own paragraph, so the text is chopped up at the width of the original page.
- Page numbers and running headers ("Chapter 3 · The Title", "142") appear in the middle of sentences, once per page.
- Hyphenated words stay broken — "recog- nize" instead of "recognize".
- Two-column pages get mixed, reading straight across both columns line by line.
- Headings aren't recognized, so there's no usable table of contents.
- Code and preformatted text lose their line breaks and indentation.
What "keeping the formatting" means in an EPUB
An EPUB is meant to reflow: the reader chooses the font and size, and the text adapts to the screen. So a good conversion doesn't copy the PDF's page design — it keeps what matters for reading:
- the chapters and sections, as a table of contents,
- paragraphs and lists,
- bold, italics and superscripts,
- code blocks with their indentation,
- images and diagrams in the right place,
- footnotes, collected at the end of each chapter.
If you truly need the exact page layout (a magazine, a comic, a form), keep the PDF — see EPUB vs PDF.
How Epubfy keeps it clean
Epubfy reads every page before converting anything:
- Layout analysis. It measures font sizes and styles, detects multi-column pages and finds text that repeats at the top and bottom of pages (headers, footers, page numbers).
- Structure detection. AI looks at samples of the layout and decides which styles are chapter and section headings. The plan is checked against the whole document, and anything that doesn't fit is corrected or dropped.
- Exact text copy. The text is copied word for word into paragraphs, with hyphenated line breaks rejoined. AI never writes or rewrites any text.
Before downloading, tap Read in browser to check the result. If the chapters look wrong, switch to basic mode, which uses font size alone to find headings.
Getting the best result
- Start from the original PDF, not a printed-and-scanned copy. If you can't select text in it, it's a scan.
- Use the full document. Repeating headers and footers are found by comparing pages, so very short excerpts give the converter less to work with.
- Fix the title on the result screen — it's what your e-reader shows in the library.