EPUB to PDF
Convert an e-book to PDF, choosing the page size that creates its pages.
This tool runs entirely in your browser. Your file is never uploaded, and you can verify that in your browser's network tab. Verify it yourself: open your browser's network tab and watch. You will see one small request asking whether you have tasks left - a tool name and a hash, never the file.
What this does
An EPUB has no pages. It is a zip of XHTML that a reader re-flows to whatever screen it is on, which is why the same book is 300 pages on a phone and 180 on a tablet. Converting one to PDF therefore means creating the pagination rather than reproducing it, and the page size you choose is not cosmetic - it decides what the book becomes. A5 is the default because it is close to a paperback; a novel set on A4 with normal margins gives lines long enough to be genuinely tiring to read.
Use it when the book has to go somewhere an EPUB cannot follow: a reader or a printer that only accepts PDF, a course pack that has to be handed round as a single file, or an annotation app that expects fixed pages to write on.
How it works
- Drop the .epub onto this page. One book at a time, up to 200 MB.
- Choose the page size. A5 is the default and reads like a paperback; A4 and Letter suit printing; A3 and Legal are there for the unusual cases.
- Set the type size between 8 and 16 points. Together with the page size and margins, this is what fixes the page count.
- Leave chapter breaks on so each chapter starts on a fresh page, and bookmarks from headings on so the outline is built from the book's own headings.
- Press Convert. The PDF downloads on its own once the whole book has been laid out.
The page count is something you create, not something you discover. It is the product of the page size, the type size and the margins, so the same book comes out at very different lengths depending on what you pick. A5 at 11 pt is close to a printed paperback. A4 at 11 pt with normal margins produces long lines and fewer, denser pages, which is fine for reference and tiring for a novel. If the length matters, try one setting, look at the result, and adjust once.
Chapter breaks decide whether the book has structure or just flow. With them on, each chapter starts on a fresh page, which is what you want for anything being printed or bound. Turn them off and the chapters run continuously into one another - a shorter document with no half-empty pages, which suits reading on a screen where nobody is turning anything over.
A fixed-layout EPUB is outside what this conversion can do faithfully. The document model behind every conversion here describes flowing content - paragraphs, headings, lists, tables, images - and has no notion of absolute positioning, so a comic, a children's picture book or a magazine whose text is placed at particular coordinates will convert, but as flowing text rather than as the layout it was. A re-flowable EPUB, which is nearly all of them, converts as intended.
The book's own fonts are not used. Text is set in the embedded Noto set instead, subsetted to the characters the book actually contains. There is a wrinkle worth knowing: the layout asks for a serif, and there is no Noto Serif in that set, so with embedding on the serif borrows the sans faces. Turning embedding off gives you a real Times Roman at the cost of every character outside WinAnsi, which for a book with accented names is usually the worse trade.
DRM is a hard stop, and this tool does not attempt to get past it and will not. A book bought from a shop that applies DRM has its chapters encrypted with a key tied to the reader it was sold for: the container still parses, and the chapters come back empty. When that happens the tool says the book is DRM-protected rather than failing with a parse error that tells you nothing.
What this tool cannot do
- DRM-protected books cannot be converted. The container parses but the chapters are encrypted, and removing that protection is not something this tool does.
- Fixed-layout EPUBs lose their layout. Comics, picture books and magazines convert as flowing text, because the model behind every conversion here has no absolute positioning.
- Fonts the book ships are not used. Text is set in the embedded Noto faces, which do not cover Chinese, Japanese, Korean, Arabic or Indic scripts.
Questions people ask
- My book will not convert. Is it the DRM?
- Almost certainly. A DRM-protected EPUB opens as a valid zip and its chapters are encrypted, so the conversion finds no readable content and tells you so. This tool does not remove DRM and is not going to. Books you wrote, books downloaded from a publisher without protection, and library exports that are DRM-free all convert normally.
- How many pages will the PDF have?
- That depends entirely on what you choose here, because the EPUB does not have a page count of its own. Page size, type size and margins together decide it. As a rough guide, A5 at 11 pt lands near a printed paperback, and moving to A4 at the same type size cuts the count substantially while making every line longer.
- Which page size should I pick?
- A5 for reading, on paper or on a phone, because the line length is close to a paperback's and that is what long reading is comfortable at. A4 or Letter if it is going to be printed on ordinary office paper or bound with other documents. Larger sizes give longer lines, so raise the type size with them or the text gets tiring.
- Do the illustrations come through?
- Images embedded in the book are carried into the flow and placed at their stated size, not re-encoded on the way. A picture used several times in the book is stored once in the PDF rather than repeated, which keeps an illustrated book from growing more than it needs to. Artwork that depends on being at a particular position on a fixed-layout page is the exception, since the layout is flow-based.
- Is the book uploaded anywhere?
- No. The EPUB is unzipped, its XHTML parsed and the PDF written entirely in a Web Worker inside your browser. Nothing about the book, including its title, leaves your device. What crosses the network is the Noto font from this site's own origin, cached after the first use, and the small request that asks whether you have tasks left - a tool name and a hash.