EPUB Find and Replace

How Text Is Stored Inside an EPUB File

An EPUB is not one long text document. It is a ZIP package containing separate HTML or XHTML chapter files plus an OPF package document, navigation files, style sheets, cover art, images, fonts, and metadata. The reading order is normally described by the package spine. A book with 30 chapters may therefore contain 30 or more searchable documents even though a reader presents them as one continuous book.

This matters because a basic ZIP-wide replacement can change far more than the words a reader sees. The same phrase might appear inside a filename, link target, CSS class, image description, or package identifier. Replacing every raw byte occurrence could break an internal link or make an XHTML document invalid. EPUB Find and Replace separates readable chapter content from the package files so the default search remains focused on visible text.

Visible Text Mode vs Raw XHTML Source Mode

Visible text mode is the safer choice for spelling corrections, character-name changes, updated terminology, and other editorial work. It parses each chapter and searches text nodes inside the document body. Element names, attributes, CSS classes, and link destinations are not included in the search. A phrase inside a paragraph can be replaced without changing the surrounding paragraph, heading, emphasis, or link elements.

Text is sometimes divided by inline markup. For example, the source may contain John <em>Smith</em> even though the reader displays “John Smith.” The tool joins adjacent text nodes for searching, so this phrase can still be found. When a selected match crosses inline elements, the replacement remains valid XHTML but may consolidate the formatting around that phrase. Those matches are marked with a warning so you can decide whether to keep them selected.

Raw XHTML source mode searches the complete source string. Use it only when the intended target is markup, an attribute, or a URL that visible mode deliberately ignores. A raw replacement can change closing tags, namespace declarations, manifest values, or link targets. Preview every raw match and make smaller, specific replacements rather than applying a broad term across the book.

How to Replace Text Across an Entire EPUB Safely

1. Keep a copy of the original EPUB before making a bulk edit.
2. Upload the book and confirm that the expected number of chapters is detected.
3. Enter the exact word or phrase to find and add the replacement text.
4. Start with Match case, Whole word, and Regular expression turned off.
5. Review the match count and open each affected chapter.
6. Deselect dialogue, quotations, names, or other occurrences that should remain unchanged.
7. Download the edited copy and open it in an EPUB reader before replacing your working file.

Whole word is useful when a short term also appears inside longer words. Match case helps separate a proper name from the same letters used as a common word. If only part of the book should change, switch to selected chapters before running the search. Navigation and package metadata remain outside the default scope and can be enabled independently under Advanced options.

Leaving Replace with empty removes the selected occurrences. The preview labels the replacement as removed, making this safer than silently deleting every match. This is useful for repeated watermarks, obsolete notices, or unwanted boilerplate, but surrounding spaces and punctuation should still be checked in each snippet.

Regex Find and Replace in EPUB Chapters

Regular expressions help when the text follows a pattern rather than one exact phrase. To change headings such as “Chapter 12” into “Part 12,” search for Chapter (\d+) and replace it with Part $1. The parentheses capture the number, and $1 inserts that captured value into the replacement. Named capture groups and $& for the complete match are also supported.

Regex uses JavaScript-compatible syntax and is never enabled automatically. Invalid expressions show an error instead of starting a search. Patterns that can match an empty string are blocked because they could create a very large number of meaningless replacements. Test a pattern on a few selected chapters before using it across the entire EPUB.

What the Rebuilt EPUB Preserves

Only files containing selected matches are rewritten. The new package copies the original folder structure, OPF package, manifest, spine, navigation, CSS, images, fonts, and other binary assets. The EPUB mimetype entry is written first without ZIP compression, while the remaining entries are compressed normally. This packaging detail helps the output remain compatible with common EPUB readers.

Before the download is offered, the browser opens the generated archive again and checks its CRC, EPUB mimetype, META-INF/container.xml, and referenced OPF package document. This is a structural package check rather than a complete EPUBCheck conformance report. It catches missing core files and failed ZIP generation, but it cannot guarantee that a risky raw-source edit still follows every EPUB or XHTML rule.

Verify the Edited Book Before Publishing

Open the downloaded file in the same reader or publishing workflow used for the original. Check the table of contents, move between several chapters, inspect images, and search for both the old and new phrases. Pay special attention to replacements near italics, bold text, footnotes, links, and chapter headings. Keep the original file until the edited version has passed those checks.

If the book does not open, return to the original EPUB, use visible text mode, and apply a smaller set of replacements. Encrypted or DRM-protected books cannot be processed reliably because their chapter content is not available as ordinary XHTML. Very large books can also use substantial browser memory; the tool limits input to 100 MB to reduce the risk of a frozen tab.