PDF to EPUB with Real Chapters & a Working TOC

Most PDF-to-EPUB tools dump the whole book into a single file and call it a day — the reader's table of contents shows one entry. This page auto-detects chapter headings, splits the book at those boundaries, and builds the NCX so your e-reader's TOC sidebar actually navigates. If auto-detect misses, the manual fallback splits every N pages so you still get jump points.

Scanned PDF? Headings can't be detected in image-only PDFs. Run Scanned PDF to Text or Make PDF Searchable first.

Drop your PDF here

or click to browse

What "chapters in EPUB" actually means

An EPUB is a ZIP that contains XHTML files (one or more per chapter), a manifest (content.opf), and a navigation file (toc.ncx in older readers, nav.xhtml in modern ones). The TOC sidebar on your Kindle, Kobo, Boox, or reMarkable reads the navigation file, not the chapter HTML. So "real chapters" means two things: the body is split into separate files (which gives you fast page turns and clean section breaks), and each chapter is registered in the NCX with the right anchor.

How the auto-detector works

After pdf.js extracts text from the PDF, the converter scans for lines that look like chapter heads. It weighs:

Anything matching becomes a chapter boundary. The chapter title goes into the NCX. The body chunk between two boundaries becomes one HTML file.

When auto-detect misses

Some PDFs are stylistic: chapter heads in tiny caps, decorative initials, no "Chapter" word anywhere. In those cases the detector finds nothing and the EPUB comes out as a single chapter. The fix is the manual fallback dropdown above — pick "every 20 pages" and you'll at least have 15 jumpable sections in a 300-page book.

What the resulting EPUB looks like in popular readers

What doesn't transfer

Sub-chapters (h2 within a chapter) get inlined as bold text rather than nested TOC entries — most readers don't render nested EPUB navigation reliably anyway. Footnotes appear inline, not as pop-ups. Page numbers from the PDF don't carry over (EPUB doesn't have fixed pages).

Privacy

The PDF is parsed locally; the EPUB is built locally. The book itself never leaves the tab. Open DevTools' Network panel during conversion to confirm.