PDF to EPUB table of contents: how it's built and how to fix a missing one

Last updated: 6 November 2026

A PDF-to-EPUB converter builds a table of contents by looking for chapter headings in the PDF, either from its built-in bookmarks or from text that looks like a heading, and then writing them into the EPUB's navigation file. If your converted book has no clickable chapters, detection found nothing it trusted, usually because the PDF has no bookmarks and its headings look like ordinary text, or because it's a scan with no real text at all.

How does a PDF-to-EPUB converter find chapters?

Converters use three signals: the PDF's own bookmarks (outline), heading-like formatting, and text patterns such as "Chapter 7". A PDF doesn't store "this is a chapter title" the way a Word file does. It stores characters at positions on a page, so structure has to be inferred.

Signal 1: the PDF's bookmarks

Many PDFs carry an outline, the bookmark panel in a PDF viewer. When it exists, it's the most reliable source because the author or publisher wrote it. When it doesn't exist, the converter has to guess. One long-running forum complaint is that a PDF's clickable contents page turned into dead links after conversion, sending readers to the cover (see this MobileRead thread). The printed contents page is just text with links to page numbers, and page numbers mean nothing in reflowable text.

Signal 2: heading formatting

Chapter titles are usually larger, bolder, centered, or followed by extra white space. A rule-based engine measures font sizes across the document, treats the largest recurring sizes as heading levels, and marks those lines as headings. This works well for novels, where every chapter opens the same way. It works badly when the title is the same size as body text, set in small caps, or placed in an image.

Signal 3: text patterns

Lines starting with "Chapter", "Part", "Book", a roman numeral or a number on its own line are strong candidates. Calibre's documentation describes the same idea: it uses XPath rules to find chapter boundaries, and you may need to adjust the XPath if no TOC is created automatically. Pattern rules misfire on repeated running headers ("MOBY DICK" at the top of every page) and on cross-references like "see Chapter 4".

Why did my converted EPUB end up with no table of contents?

Usually one of four causes: the PDF is scanned, the headings aren't visually distinct, running headers confused the detector, or the book's contents page was converted as plain text rather than navigation. Each leaves a different symptom.

If you're unsure whether your file is scanned, our 60-second check for scanned vs text PDFs takes care of the first row.

What makes an EPUB table of contents actually clickable?

An EPUB's clickable contents comes from its navigation document, a special XHTML file with a list of links to chapter files. The EPUB 3.3 specification says a valid navigation document must include exactly one toc nav element. The older NCX file was the previous way of doing this and is now considered a legacy feature, per the EPUB 3.2 spec.

This matters because a "Contents" page typed into the book body is not the same thing. Your reader's menu, the Kindle's "Go to" list and Apple Books' chapter list all read the navigation document, not the page text. So an EPUB can look like it has a contents page and still show nothing in the reader's menu.

A good navigation file also has these traits:

• One entry per real chapter, in reading order.
• Each link points to the start of a chapter, not the middle of the previous one.
• Nested levels (Part, then Chapter) when the book has them.
• Titles that match the headings in the text.

How do I fix a missing table of contents? (Steps)

Work from the cheapest fix to the most involved.

• Check the source PDF first. Open it and look for a bookmarks panel. If there is one, the structure exists and a decent converter should use it. If not, scroll through and note how chapters begin: big titles, "Chapter" labels, or just a new page.
• Check whether it's scanned. Try selecting a sentence. If you can't, the file is images, and you need OCR before any chapter can be detected. Our guide to converting a scanned PDF to EPUB with OCR walks through it.
• Convert with chapter detection on. Don't pick a fixed-layout mode for a novel. Fixed layout keeps each page as a picture-like unit, which defeats text-based structure.
• Open the result and test the menu, not just the pages. In your reader, open the contents list and tap three entries, one near the start, one in the middle, one near the end.
• If entries are wrong, fix the source rather than the EPUB. Delete repeated headers and footers, or retype chapter headings in a Word or text copy with proper heading styles, then convert that. Formats with real heading styles carry structure across cleanly.
• Last resort: edit the EPUB. Tools like Calibre let you define chapter-detection XPath rules or build a TOC by hand, which is powerful but fiddly. We cover this in our fix for an EPUB with no table of contents.

What about textbooks and technical PDFs?

These are the hardest. Multi-column layouts, sidebars, figure captions and equations all look like headings or break the reading order. A simple rule-based pass copes with novels but can mis-detect here. Convert2EPUB offers a free Standard model that it describes as good for simple layouts like novels, and a Pro AI option (2 credits per page) for more complex documents. We'd recommend testing a short section of a complex book before converting the whole thing.

Does the original PDF's quality change the result?

Yes, more than any converter setting does. Detection can only use what the file contains, so a tidy PDF exported from a word processor with real heading styles gives a better table of contents than a scan of a printed book.

Two things to be realistic about. First, no converter can recover chapters the file doesn't signal in any way, such as a chapter that begins with only a small paragraph break. Second, OCR errors in a heading ("Chapfer 3") can defeat pattern matching, so skim your headings after an OCR conversion.

Convert2EPUB's converter detects chapters automatically and builds a working table of contents, and its OCR option turns scanned pages into real text first. If a book still comes out without chapters, you can convert a PDF to EPUB for free and compare Standard and Pro results on a short sample.

Frequently asked questions

Can I keep the clickable contents page from my PDF?

Not as it is. The PDF's page-number links don't map onto reflowable text, so they usually break. The better route is to have the converter regenerate the contents from detected headings, which gives links that work in any reader.

Why does my Kindle show no chapters from my EPUB?

The EPUB probably has no navigation document, or it lists only the title. Typed contents pages inside the book won't appear in the Kindle's menu. Check the EPUB in another reader to confirm, then regenerate the TOC.

Does a fixed-layout EPUB have a table of contents?

It can, since the navigation document works the same way, but detection is harder because each page is treated as a unit. For text-heavy books, reflowable output gives better chapter structure and readable text.

Do I need Calibre to fix this?

No, but it's a common option. Calibre exposes chapter-detection rules and a manual TOC editor, which suit people comfortable with XPath. Starting with a cleaner source file or a converter with automatic detection is usually quicker.