PDF to EPUB Conversion Problems — Troubleshooting Guide (2026)

Most PDF to EPUB conversions work without any issues. But some PDFs have quirks that cause problems: garbled text, missing images, wrong column order, or unexpected formatting. This guide covers the most common issues and what to do about each one.

Problem: Text is garbled, scrambled, or contains strange characters

Cause: The PDF uses an embedded font with a non-standard character encoding. The PDF displays correctly because the rendering engine knows the font, but text extraction produces incorrect characters.

Common sources: PDFs created by older publishing tools (especially pre-2010), PDFs with embedded LaTeX fonts, PDFs from certain Asian-language publishing systems.

Fixes to try:

1. Download the PDF directly from the original publisher — cleaner source files have standard encoding
2. For academic papers, try the arXiv, PubMed, or institutional repository version instead of a third-party download
3. If the PDF is a scan (not a native digital document), use the scanned PDF workflow which applies OCR rather than text extraction

Problem: Images or figures are missing from the EPUB

Cause: Images may be embedded as vector graphics (PDF vector objects) that weren't rasterized, in CMYK color space, or using non-standard compression.

Fixes to try:

1. Download the PDF directly from the publisher — third-party re-uploads sometimes strip or re-compress images
2. Check whether the images are present in the PDF itself by opening it on your computer — if they appear in the PDF, they should appear in the EPUB
3. Very small thumbnails in the PDF will produce small images in the EPUB — the converter doesn't upscale

Problem: Reading order is wrong (columns mixed up)

Cause: The converter reads across the page instead of down each column. This happens when column detection fails on unusual layouts.

Fixes to try:

1. Try the original publisher PDF rather than a re-uploaded version — cleaner layout metadata helps column detection
2. For arXiv papers, try the PDF version directly from arXiv rather than a third-party mirror
3. If the PDF has a mixed layout (e.g., one-column abstract + two-column body), some columns may still read correctly

Problem: EPUB looks different in different readers

Cause: EPUB readers have significant variation in how they render CSS, handle fonts, and display tables. This is expected behavior — not a conversion bug.

Notes:

• Apple Books tends to render the most faithfully to the source formatting
• Kindle overrides fonts and some CSS — this is by design and improves readability on e-ink
• KOReader on BOOX or reMarkable has its own typography engine
• In all cases, the text content is identical — only visual presentation differs

Problem: Conversion fails or hangs

Cause: Very large PDFs (over 50 MB), PDFs with unusual internal structure, or corrupted PDF files.

Fixes to try:

1. Check the file size — the free tier handles up to 50 MB
2. Try re-downloading the PDF — the file may be corrupted or incomplete
3. For very long documents, consider splitting the PDF by chapter and converting each part separately
4. Open the PDF on your computer first to confirm it isn't password-protected or DRM-locked (DRM-protected PDFs cannot be converted by any tool)

Frequently asked questions

Why does my converted EPUB have garbled or scrambled text?

Garbled text usually means the PDF uses a non-standard or embedded font where character encoding is not mapped correctly. This is common in older PDFs or PDFs created by non-standard tools. Try re-downloading the PDF from the original source — publisher PDFs are often cleaner than third-party re-uploads. If the original PDF is the only source, OCR-based conversion (treating the PDF as a scan) may produce better text extraction.

Why is my EPUB missing images?

Images may be missing if they were embedded as vector graphics that couldn't be rasterized, or if they are in an unsupported format. Some PDFs embed images in unusual color spaces (CMYK, spot colors) that require conversion. Try a different PDF — the same paper downloaded from the publisher's site is often cleaner than a third-party scan. If images are present but very small, the original PDF may have compressed them aggressively.

Why is the reading order wrong in my EPUB?

Wrong reading order is almost always caused by two-column or multi-column PDFs where the converter reads across both columns instead of down each column first. The toolkit.bot converter has dedicated two-column detection, so this should be rare. If it happens, check whether the PDF is actually a two-column layout — some complex PDFs have mixed layouts (e.g. title page single-column, body two-column) that can confuse column detection.

Why does my EPUB look fine in one reader but broken in another?

EPUB readers vary significantly in how strictly they interpret the EPUB spec. A rendering that looks fine in Apple Books may look different in Kindle or KOReader. Common issues: (1) Fonts — some readers ignore embedded fonts. (2) CSS — some readers apply their own stylesheets that override the EPUB's. (3) Tables — complex tables may reflow differently. For academic papers, this is usually minor and the text content is still fully readable.

My EPUB file is much smaller than the PDF — did something get lost?

Not necessarily. EPUB compresses text very efficiently because text is stored as HTML, not as rendered glyphs. A 2 MB PDF may produce a 200 KB EPUB with identical text content. Images are the main factor in EPUB file size — if your EPUB seems very small and the PDF was image-heavy, check that figures are present in the EPUB before assuming they were lost.