EPUB vs PDF: Which eBook Format to Convert To, and What Breaks

An EPUB is a zip of HTML, CSS and assets with a manifest. The text is text, so the reader controls font size, typeface, margins and line spacing, and the content reflows to the screen. That is ideal for novels, essays and anything read linearly on a phone, and it is what accessibility tools depend on. It also means the author cannot control exactly where a page breaks or how a line falls.

A PDF is a fixed page. Every element has a coordinate and a size, so a two-column layout with captions, sidebars and precise typography looks the same on every device. That is ideal for manuals, textbooks, reports and anything where layout carries meaning. The cost is that reading a fixed A4 page on a phone means zooming and panning, and text extraction depends on how the PDF was produced.

A PDF does not know what a paragraph is. It knows where glyphs were drawn. A converter has to infer structure from positions: lines close together form a paragraph, a short line followed by a gap may be a heading, a repeated footer must be removed. When the inference is right the result reads well; when it is wrong you get broken sentences, headings buried in body text, and page numbers scattered through the prose.

Images are extracted as separate files and may lose their position relative to the text, tables often collapse into unreadable sequences of cells, and multi-column layouts are the hardest case. Scanned PDFs are worse still: there is no text at all, only pictures of words, so conversion produces either images of pages or garbage unless OCR runs first. Expect to proofread and repair, and do not promise a clean automatic result.

This direction is more predictable. The EPUB's HTML is rendered to pages, so the text, images and basic formatting come through, but the reader-controlled settings are frozen at whatever the renderer chose. Font size, page size and margins become fixed, and a book that looked comfortable on a phone may end up with very wide lines on A4 unless the page size is set sensibly.

Fonts referenced by the EPUB but not embedded in it will be substituted, which can change line breaks. Fixed-layout EPUBs, used for comics and children's books, convert much more faithfully because they already specify positions. Reflowable books with complex tables or code blocks are where the layout tends to suffer.

Also check for DRM. A protected EPUB cannot be converted by an ordinary tool, and the failure usually surfaces as an unreadable or empty output rather than a clear message. Only convert files you have the right to use.

Send EPUB to people who will read on a phone, a tablet or a dedicated e-reader: the text scales, the file is small, and the reading experience is what those devices are built for. Send PDF when the layout is the message, when the document will be printed, or when the recipient needs a file that looks the same on every machine. Sending both is a legitimate answer and often the kindest one.

When converting, keep the source. A PDF to EPUB conversion is a lossy interpretation of structure, not a reversible transformation, so going back rarely recovers the original layout. If the book is still being edited, work from the original manuscript format — DOCX, Markdown or HTML — and generate both outputs from it rather than converting one output into the other.