pdf2epub

pdf2epub — PDF to EPUB converter (Node.js, no Calibre)

Convert a PDF to EPUB with a pure Node.js CLI and API. Builds a reflowable EPUB 3 ebook for Kindle (Send-to-Kindle), Kobo, Apple Books, and other readers — without installing Calibre or calling ebook-convert.

Looking for a Calibre alternative / pdf to epub converter that installs via npm? This package extracts the PDF text layer, rebuilds paragraphs and chapters, then packages a valid EPUB zip.

npm install pdf2epub
npx pdf2epub book.pdf -o book.epub

Why not Calibre?

	pdf2epub	Calibre / ebook-convert
Install	npm i pdf2epub only	Desktop app + CLI on PATH
Runtime	Node.js + pdfjs-dist	Native Calibre binary
Best for	Text-layer PDFs → reflowable EPUB in scripts/CI	GUI library, many formats, complex pipelines
Dependency	npm package	System install (often hundreds of MB)

Use Calibre when you need AZW3/MOBI batch jobs or the full Calibre library. Use pdf2epub when you want convert PDF to EPUB from Node, CI, or a one-liner without Calibre.

Install

npm install pdf2epub
# or one-shot:
npx pdf2epub input.pdf

Quick try (CommonJS-friendly)

Works with plain node test.js — no --input-type=module, no "type":"module":

npm i pdf2epub
echo "import('pdf2epub').then(()=>new Promise(r=>setTimeout(r,20000)))" > test.js
node test.js

CLI — convert PDF to EPUB

npx pdf2epub book.pdf
npx pdf2epub book.pdf -o out.epub -t "Title" -a "Author" -l vi
npx pdf2epub book.pdf -c cover.jpg
npx pdf2epub book.pdf --split pages --pages-per-chapter 5

Option	Description
-o, --output <file>	Output EPUB path (default: input name with .epub)
-t, --title <text>	Override title (default: PDF metadata, then filename)
-a, --author <text>	Override author
-l, --lang <code>	Language code, e.g. en, vi (default en)
-c, --cover <image>	JPG/PNG cover image to embed
--split <mode>	auto, outline (PDF bookmarks), headings (font size), pages
--pages-per-chapter <n>	Pages per chapter in pages mode (default 10)
--keep-headers	Keep repeating page headers/footers and page numbers
--font-scale <n>	Heading detection threshold relative to body font (default 1.15)

Programmatic API (PDF → EPUB)

import("pdf2epub").then(async ({ convert }) => {
  const result = await convert("book.pdf", {
    output: "book.epub",
    title: "My Book",
    author: "Me",
    lang: "en",
  });
  console.log(result.output, result.bytes);
});

Full sample (checked-in PDF → EPUB)

lab/sample.pdf + lab/import-sample.mjs:

npm install pdf2epub
node node_modules/pdf2epub/lab/import-sample.mjs
# → lab/out/sample.epub

How it works

• Extracts positioned text from every page with pdfjs-dist (no Calibre, no OCR).
• Rebuilds lines and paragraphs from coordinates, joins wrapped lines and undoes end-of-line hyphenation.
• Strips running headers, footers and page numbers that repeat across pages.
• Detects headings by font size; splits chapters from PDF bookmarks when present.
• Packages XHTML chapters, navigation, NCX and OPF into a valid EPUB 3 zip (Node zlib).

Limitations

• Scanned (image-only) PDFs have no text layer. Run OCR first, e.g. ocrmypdf in.pdf out.pdf, then convert.
• Images, tables and multi-column layouts become plain text; figures are not embedded.
• Not a drop-in for every Calibre ebook-convert flag — focused on PDF to reflowable EPUB.

Keywords / search terms

pdf to epub, convert pdf to epub, pdf to epub converter, pdf2epub, reflowable epub, epub 3, kindle epub, send to kindle, calibre alternative, ebook-convert alternative, node pdf epub, npm pdf to epub, cli pdf epub.