Extract Images from EPUB

An EPUB is, underneath, a ZIP archive of HTML, CSS, and media. This tool opens that archive in your browser, reads the book's manifest, and surfaces every file the publisher declared as an image so you can lift it out without unzipping anything by hand or hunting through nested OEBPS folders.

It fits a handful of real jobs. Self-publishers and editors recover a cover, chapter ornaments, or diagrams when the original source files have been lost and only the compiled EPUB remains. Designers and developers grab illustrations to reuse on a landing page, a sample chapter, or a store listing. Students and researchers pull figures, maps, and charts out of a reference title for a citation or a slide.

It is also a quick archival step: when you want a clean, flat folder of just the pictures from a book, selecting everything and exporting a ZIP gives you exactly that, separated from the markup around it.

Drag an .epub onto the dropzone or click to select one. The tool parses the archive locally, scans the manifest for entries with an image media type (JPEG, PNG, GIF, SVG, WebP, and similar), reads each one from the ZIP, and builds a thumbnail grid. A summary across the top shows the total image count, how many are currently selected, and the combined size of your selection.

Every image is selected by default. Each tile previews the picture and labels it with its filename and individual file size; tap a tile to toggle it, or use Select all and None to adjust the whole set at once. When you are ready, the button exports your selection: a single picked image downloads on its own under its original filename, while two or more are bundled into a ZIP named after the book.

Two honest notes on scope. The tool reads images the EPUB's manifest declares, so a stray picture that the book never references may not appear. And while it reports filenames and byte sizes, it does not display pixel dimensions in the grid, despite that being a natural thing to expect; check resolution in your image viewer after export if it matters.

DRM is a hard wall. If the file is encrypted by a store's protection scheme, the tool detects it and refuses rather than producing scrambled output, because the image bytes are not readable. Library loans and many purchased store copies fall into this category. A book that genuinely contains no images will also report nothing found rather than an empty download.

The most common surprise is fonts and styling. This extracts raster and vector images only, not embedded fonts, CSS, or the text itself; for words use a text or Markdown extractor instead. Another is filename collisions inside the ZIP. EPUBs sometimes reuse plain names like image1.jpg across different folders, and because the export flattens everything to its base filename, two such files could overwrite each other in the archive.

Finally, everything runs in the page. A very large book with hundreds of high-resolution images holds all of them in memory at once to build previews, so an unusually heavy file can feel slow or strain a low-memory device. Nothing is uploaded and nothing persists once you close the tab, which also means a refresh discards your selection.

Start with the count: does the number of files in your download match the selected total shown in the tool? Open a few thumbnails at full size to confirm they are the actual artwork and not low-resolution placeholders, and that the cover came through if you wanted it. Watch for any duplicated filenames in the ZIP, since a collision means one version quietly replaced another.

If a specific picture is missing, it may not have been declared in the manifest, or the book may be DRM-protected, in which case the tool would have told you up front. Re-check the source file in that case rather than assuming the export dropped it silently.

For anything official, financial, or legal, treat the extracted images as a draft and verify by hand. Open each file in a dedicated image viewer to confirm resolution, color, completeness, and that no figure was cropped or substituted, and reconcile the contents against the original book before relying on them in a published, submitted, or contractual document.