EPUB to PDF Converter

An EPUB is a ZIP archive containing XHTML chapter files, optional CSS, fonts, and images, plus a small set of XML manifests that describe the reading order. The W3C's EPUB 3.3 Recommendation locks down the structure: the first file in the archive is the uncompressed text application/epub+zip (the OCF mimetype sentinel), then META-INF/container.xml points at the OPF package document, and the package document's spine declares the chapter order. This tool walks that structure deterministically:

1. Read the ZIP central directory. A handwritten parser locates the End-of-Central-Directory record by scanning backward from the end of the file (PKWARE APPNOTE.TXT §4.3.16), then walks each Central Directory File Header to enumerate the entries. ZIP64 containers are flagged with a friendly error instead of silently mis-parsed.
2. Decompress entries on demand. The browser's native DecompressionStream("deflate-raw") handles compression method 8 (DEFLATE) — the only compressed method EPUBs use. Stored entries (method 0, including the mimetype sentinel) are copied byte-for-byte. No third-party ZIP library is loaded.
3. Parse the OPF spine. The package document's <manifest> lists every resource in the book; the <spine> lists the reading order. Items with linear="no" are excluded — those are typically pop-up notes the reader does not want printed.
4. Stitch the chapters into one document. Each chapter's body content is concatenated into a single HTML document with a CSS Paged Media @page rule for the chosen size and margin. Scripts, custom CSS, and event-handler attributes are stripped so nothing executes inside the sandboxed preview iframe; a curated print stylesheet replaces them so the output is consistent across books.
5. Print to PDF natively. The assembled document is rendered in a sandboxed same-origin iframe, and iframe.contentWindow.print() opens the browser's print dialog. Picking Save as PDF writes the PDF directly to disk — selectable text, working hyperlinks, no rasterisation.

Two independent estimators compute the expected page count: a layout formula based on the chosen page size, font size, and margin, and a density formula tuned to the canonical 250-words-per-A5-page ratio that publishers use for paperback planning. Both numbers are shown above. When they disagree by more than 30%, the UI flags it so you can sanity-check before printing.

• W3C — EPUB 3.3 (Recommendation, 25 May 2023): package document, OCF container, spine
• W3C — EPUB 3.3 §6 Open Container Format (mimetype sentinel rules)
• PKWARE — .ZIP File Format Specification (APPNOTE.TXT 6.3.10)
• W3C — CSS Paged Media Module Level 3 (Working Draft) — @page rules and margins
• MDN — DecompressionStream API (deflate-raw token for raw DEFLATE streams)
• ISO 216 — A-series paper sizes (A4, A5, B5)

Specifications and the dependency list were last cross-checked on 2026-05-11. The page is re-verified annually or whenever EPUB ships a minor revision — whichever is sooner.