EPUB to PDF Converter

An EPUB is fundamentally a ZIP archive that follows the Open Packaging Format (OPF) specification. Its internal structure is fixed: the first entry is an uncompressed mimetype file containing application/epub+zip (for quick format identification); next is META-INF/container.xml, which points to the root .opf file; the .opf file has three sections - metadata (describing title, author, language, and a unique identifier using Dublin Core elements), manifest (listing every resource in the package with its properties), and spine (defining the reading order of content documents). The body consists of XHTML files (OPS 2.0 for EPUB 2, an HTML5 subset for EPUB 3), styled with CSS, with images and fonts stored as separate resources; navigation is provided by NCX (toc.ncx) in EPUB 2 and nav.xhtml in EPUB 3. EPUB 3 additionally supports MathML, SVG, audio/video, scripting, and fixed-layout (FXL) documents declared via the rendition:layout=pre-paginated property with explicit page dimensions.

The core difficulty in EPUB-to-PDF conversion is 'reflow to pagination'. EPUB is reflowable by default - text flows and wraps based on screen width and the user's font-size setting, with no concept of a 'page'; PDF is fixed-layout, where every element has absolute coordinates. The converter therefore needs a built-in layout and pagination engine: it reads the XHTML body and CSS, computes line boxes according to the chosen page size (e.g. A4, Letter) and margins, fills text line by line until it nears the page bottom, then decides page breaks based on the CSS page-break-before / page-break-after / break-* properties and widow/orphan control on headings, slicing the flowing content into fixed pages. This process closely mirrors browser rendering plus print pagination, so many converters render the XHTML in a headless browser (headless Chromium) and call print-to-PDF for the highest CSS fidelity; Calibre's ebook-convert uses its own rendering pipeline with output profiles and page-size parameters; Pandoc takes the EPUB -> intermediate format -> LaTeX -> PDF (xelatex) route, which works well for text-heavy books but has limited CSS fidelity.

Font handling is the key to fidelity. EPUB references embedded .ttf/.otf fonts in the fonts/ directory through @font-face; when converting to PDF, these fonts must be embedded as subsets (PDF Type 0 CID-keyed fonts with an embedded TrueType program, per PDF spec section 9.6.5), otherwise the PDF will repaginate using a system substitute font and every line break and page break will shift. Content in Chinese, Japanese, Arabic, and other scripts that require complex text shaping (ligatures, variant selectors, writing direction via HarfBuzz) is especially dependent on correct font embedding. Image streams are recompressed: raster images go through DCT (JPEG) or Flate filters depending on transparency, and SVG vector graphics are converted to PDF content-stream operators (m/l/c/re/S/f) to stay scalable or are rasterized outright. The cover image is typically rendered as the first page. Fixed-layout EPUBs (FXL, such as picture books and comics) have explicit page dimensions, so they map to PDF most directly - almost one page to one page.

This tool only accepts DRM-free EPUB files. Store-bought titles protected by Adobe ADEPT, Apple FairPlay, or other DRM schemes cannot be parsed; .mobi and .azw3 are Kindle-proprietary formats, not EPUB, and require other tools. The uploaded EPUB is processed by the server-side conversion engine and deleted from the server immediately after the PDF is generated. The output follows the ISO 32000-1:2008 (PDF 1.7) or ISO 32000-2:2020 (PDF 2.0) specification.

• EPUB container structure: the ZIP holds an uncompressed mimetype (application/epub+zip), META-INF/container.xml pointing to the root .opf, and the .opf with its metadata (Dublin Core) / manifest / spine sections
• Content documents: EPUB 2 uses OPS 2.0 XHTML, EPUB 3 uses an HTML5 subset with CSS3, MathML, SVG, audio/video, and scripting support; navigation is NCX (EPUB 2) or nav.xhtml (EPUB 3)
• Reflow to pagination: EPUB is reflowable (text flows and wraps), PDF is fixed-layout; the converter's pagination engine slices pages based on page size, margins, font size, and CSS page-break / break-* properties
• Conversion engines: headless Chromium print-to-PDF (highest CSS fidelity), Calibre ebook-convert (own rendering pipeline + output profiles), Pandoc (via LaTeX, good for text-heavy books)
• Font fidelity: @font-face embedded fonts must be embedded as TrueType subsets in the PDF (Type 0 CID, spec 9.6.5); missing fonts trigger a system fallback and repagination, while CJK and complex scripts depend on HarfBuzz shaping
• DRM and format limits: only DRM-free EPUB is accepted; Adobe ADEPT / Apple FairPlay titles cannot be parsed; .mobi / .azw3 are Kindle-proprietary, not EPUB
• Files are processed by the server-side conversion engine and deleted immediately after the PDF is generated; output follows ISO 32000-1 (PDF 1.7) or ISO 32000-2 (PDF 2.0)