EPUB to PDF and Kindle: Private Browser Conversion 2026

An EPUB file is a zipped bundle of XHTML, CSS, images, and a manifest. A PDF is a fixed-layout description of pages. A Kindle file (AZW3 for older devices, KFX for anything bought after 2017, and KPF for sideload) is its own Amazon-specific container wrapping what is essentially modified HTML. These formats do fundamentally different things, and a conversion between any two of them is a tradeoff, not a translation.

Understanding what breaks helps you pick the right output.

What actually lives inside an EPUB

Open any EPUB with unzip and you will see something like this:

META-INF/container.xml
OEBPS/content.opf
OEBPS/toc.ncx
OEBPS/nav.xhtml
OEBPS/chapter-01.xhtml
OEBPS/chapter-02.xhtml
OEBPS/styles/main.css
OEBPS/images/cover.jpg

The .xhtml files carry the text, marked up with real semantic HTML. The CSS controls typography. The manifest (content.opf) lists every resource and its media type, and the nav.xhtml provides the table of contents. EPUB 3.3, the current spec as of 2023, also allows scripting, MathML, and SVG with full CSS3 support.

The critical property of EPUB is that it is reflowable. The reader picks the font size, the device picks the page width, and the text rewraps to fit. Page numbers are not a thing.

EPUB to PDF: pick fixed or reflowable

When you convert to PDF, you have to freeze the layout. Two reasonable approaches:

Fixed pagination at a target page size. Render the EPUB as if it were a printed book. A5 (148x210mm) matches most trade paperbacks. 6x9 inches is standard for US trade. The conversion picks a font size (11–12pt serif for body is conventional), sets margins, and paginates. Output looks like a real book PDF. Good for printing. Bad for reading on a phone because the text will not reflow.

Tall single-page layout. Render the entire book as one very long page. The reader scrolls. Bad for printing, better for phones. Almost nothing does this well, but it is worth knowing it exists.

For most uses, a fixed 6x9 with 1-inch margins and a 12pt serif is fine. The choices that actually matter:

• Preserve embedded fonts. If the EPUB embeds a font for a reason (a cookbook with a decorative display face, a comic with lettering), the PDF conversion should preserve the font instead of substituting.
• Image quality. EPUBs often contain already-compressed JPEGs. A naive conversion re-encodes them again at lower quality. Keep the originals.
• Keep the table of contents. Export it as a PDF outline (bookmarks). A 400-page book without a navigable TOC is unusable.

EPUB to Kindle: AZW3, KFX, or KPF

Amazon's ecosystem has three relevant formats.

Format	Used by	How to make one	Notes
MOBI	Deprecated, removed from Send-to-Kindle in 2022	Don't, really	Dead format
AZW3	Older Kindles, Paperwhite pre-2017	Calibre, kindlegen (also deprecated)	Works via email sideload
KFX/KPF	Everything current	Kindle Previewer 3	The modern target

Since 2022, Amazon's Send-to-Kindle service accepts EPUB directly and converts on their side. If you are happy to hand the file to Amazon, just email it to your @kindle.com address and skip the conversion entirely. If you are not, generate a KPF locally with Kindle Previewer and sideload via USB or Send-to-Kindle's file upload.

Typography-wise, Kindle rewraps text the same way EPUB does, so reflowable content survives the trip. Fixed-layout EPUBs (children's books, comics, cookbooks with rigid grids) translate to fixed-layout KFX, which looks fine on Fire tablets and unreadable on a 6-inch Kindle.

DRM: the one thing we cannot help with

If the EPUB came from a store, it is probably DRM-protected. Adobe ADEPT for most non-Amazon retailers, Amazon's own DRM for Kindle content. A DRM-wrapped file is not a valid EPUB; it is an encrypted blob with an EPUB-shaped manifest. You cannot convert it without first removing the DRM, and removing DRM from content you did not buy the license to strip is not something any legitimate tool will help with.

For DRM-free EPUBs (Project Gutenberg, Standard Ebooks, direct purchases from Tor.com, publisher direct sales, your own publications), conversion is straightforward.

Running it privately in the browser

A browser-based conversion handles EPUB end-to-end in WebAssembly: unzip, parse XHTML, apply CSS, paginate, render. The book text and personal notes stay on your device. Konvrt does this at /convert for single files, and /batch handles an entire Calibre library export without per-file reuploads.

What to expect on output size:

• A 300-page novel (pure text): EPUB 400 KB, PDF 1.2 MB, AZW3 500 KB
• A cookbook with 80 photos: EPUB 25 MB, PDF 30 MB (minor bloat from PDF font tables), AZW3 22 MB
• A comic in fixed-layout EPUB: EPUB 120 MB, PDF 110 MB, KPF 90 MB

One specific gotcha

EPUBs that use CSS page-break-before: always in odd places will produce PDFs with blank pages between chapters. If the output looks wrong, open the EPUB, check for page-break declarations in the CSS, and either strip them or let the converter respect them. Most converters default to respecting them, which is usually what you want.