PDF to EPUB — Convert PDF to EPUB in Your Browser
Convert PDF files to EPUB ebooks. Text-only reflowable output, no uploads, no accounts, runs entirely in your browser.
Drop your PDF file here
or click to browse (max 50MB) — text only, images not preserved
Text-only conversion. Images and complex layouts in the source PDF are not preserved.
PDF to EPUB is one of those conversions that sounds trivial and isn’t. The two formats are doing fundamentally different jobs: a PDF is a fixed-layout print-ready document with glyphs at exact coordinates, and an EPUB is a reflowable stream of text and styles that the reader paginates on the fly. Converting between them means either preserving the exact look (which an EPUB cannot do, by design) or rebuilding the reading flow from the glyphs (which is what this tool does, with the limits that implies).
The honest version of this tool is a text extractor that packages its output as an EPUB. It walks every page, sorts the glyphs by position, groups them into lines, and writes each line as a paragraph. There is no OCR, no layout analysis, no figure detection, no column detection. For prose PDFs — books, articles, contracts, reports — the result is readable. For designed documents — brochures, multi-column papers, anything with figures or footnotes — the result is the words in the right order, but the visual structure is gone. That trade-off is documented in the UI and in the limitations list, and it is the right call for an MVP that runs in a browser.
The next iteration, when it comes, will add an option to embed page images instead of extracting text — useful for comics, magazines, and designed layouts where reflow is the wrong output anyway. Until then, this tool is the right pick for plain-text PDFs, and the right pick is to OCR first for anything that is a scan.
How to use
Drop a PDF
Drag a .pdf file into the drop zone, or click to browse. The file is parsed in your browser using pdfjs-dist — the bytes never leave the device.
Wait for text extraction
The tool walks every page, reads the text layer, and groups glyphs into lines. PDFs without a text layer (pure scans) produce empty pages — run OCR first.
Download the EPUB
The output is a valid EPUB 3.0 file with one chapter per page, a navigation document, and an NCX table of contents for older readers.
Frequently asked
Why does the EPUB have no images?
Real image-preserving PDF to EPUB conversion requires running OCR or layout analysis, classifying figures, and repositioning them around the text. This tool ships the simpler version — text only — and documents that limit in the UI.
My PDF is a scanned image with no text. Will it work?
No. The converter reads the text layer; a pure scan has none, so the output is a list of empty pages. Run an OCR tool on the PDF first, then convert.
Why is each PDF page its own chapter?
Without running layout analysis we cannot detect chapter boundaries. Page-level chapters are predictable and let the reader navigate. Some readers will treat this as a long book of very short chapters.
Will footnotes and table cells be preserved?
Partially. The converter sorts glyphs by position and reconstructs lines, so most prose survives intact. Multi-column layouts, sidebars, and table cells may merge in the wrong order.
What reader apps will open the output?
The EPUB is EPUB 3.0 compliant with a fallback NCX, so it opens in Apple Books, Google Play Books, Kobo, Calibre, Thorium, and any other modern reader.
Limitations
- Text-only outputEmbedded images, vector graphics, charts, and tables are dropped. The EPUB contains only the text layer, rendered as one paragraph per reconstructed line.
- No layout preservationHeadings, columns, sidebars, and footers are inferred from glyph position and sometimes land in the wrong order. The output is reflowable, which is the point of EPUB, but the page boundaries from the source do not survive.
- Scanned PDFs produce empty pagesA scanned PDF without a text layer yields chapters with no text. OCR must be run first — this tool does not perform OCR.
- Large font and language support is limitedGlyphs are merged by spatial proximity. Right-to-left scripts and complex scripts (Arabic, Devanagari, Thai) may be reordered incorrectly. Latin and CJK generally work.
Platform notes
- macOS
- Apple Books opens the generated EPUB without complaint. Drag the file into Books or use File > Import. Pages with reflowed text look correct in the default serif font.
- Windows
- Microsoft Edge can read EPUB natively (drag the file into a tab). Calibre, Thorium, and Sigil all open the file; Sigil will warn about the missing images and let you inspect the spine.
- Linux
- Calibre's ebook-viewer opens the output and shows the chapter list in the sidebar. Foliate and KOReader also handle the NCX table of contents.
- iOS
- iOS Books imports the EPUB through Files > Share > Books. The chapter list appears as expected; right-to-edge taps advance between pages because each PDF page is one chapter.
- Android
- Google Play Books, ReadEra, and Lithium all open the output. Play Books shows the chapter list pulled from the NCX.