epub

Convert EPUB ↔ PDF bidirectionally. EPUB→PDF preserves layout for analysis; PDF→EPUB compresses and reflows for distribution and ereaders.

Safety Notice

This listing is from the official public ClawHub registry. Review SKILL.md and referenced scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "epub" with this command: npx skills add bunsdev/epub-to-pdf

EPUB Skill

When to use

EPUB → PDF

  • Convert EPUB books, papers, or distributed documents to PDF for reading and analysis
  • Preserve layout and formatting before text extraction via pdftoppm
  • Archive EPUB content in a fixed-layout format
  • Prepare EPUBs for downstream processing (rendering, analysis, synthesis)

PDF → EPUB

  • Convert papers and PDFs to reflowable EPUB format for ereader distribution
  • Compress PDFs into smaller, device-friendly EPUB archives
  • Auto-detect chapters and structure for better reading experience
  • Redistribute research materials in open formats

Workflow

EPUB → PDF (for analysis)

  1. Check for calibre → install if missing (brew install calibre)
  2. Convert EPUB → PDF using ebook-convert with formatting options
  3. Validate output → confirm PDF rendered correctly
  4. Render pages → use pdftoppm to extract visual content
  5. Analyze → pass to PDF skill for text extraction and synthesis

PDF → EPUB (for distribution)

  1. Check for calibre → install if missing
  2. Convert PDF → EPUB using ebook-convert with reflowable settings
  3. Auto-detect chapters → ebook-convert scans for structure
  4. Validate compression → confirm EPUB is smaller than PDF
  5. Archive → save EPUB to research/sources/ for distribution or ereader use

Usage

EPUB → PDF (basic)

ebook-convert input.epub output.pdf

EPUB → PDF (with layout preservation options)

ebook-convert input.epub output.pdf \
  --pdf-standard-font "sans" \
  --paper-size "a4" \
  --margin-left 20 \
  --margin-right 20 \
  --margin-top 20 \
  --margin-bottom 20 \
  --pdf-mono-family "mono"

Note: Valid font families are: serif, sans, mono (not Helvetica, etc.)

PDF → EPUB (basic)

ebook-convert input.pdf output.epub

PDF → EPUB (with compression)

ebook-convert input.pdf output.epub \
  --output-profile "tablet" \
  --paper-size "a4"

Note: ebook-convert auto-detects chapters from PDFs. Inspect the output EPUB to verify structure.

In context (Sage workflow)

EPUB → PDF

# Convert EPUB to PDF for analysis
ebook-convert "/path/to/input.epub" "/path/to/research/sources/output.pdf"

# Verify and check size
ls -lh "/path/to/research/sources/output.pdf"

# Render pages for visual inspection
pdftoppm -png "/path/to/research/sources/output.pdf" "/tmp/preview/page"

PDF → EPUB

# Convert PDF to EPUB for distribution
ebook-convert "/path/to/input.pdf" "/path/to/research/sources/output.epub"

# Verify compression
ls -lh "/path/to/input.pdf" "/path/to/research/sources/output.epub"

# (EPUB will typically be 30-50% of PDF size)

Dependencies

System package:

# macOS (Homebrew)
brew install calibre

# Ubuntu/Debian
sudo apt-get install -y calibre

# Fedora/RHEL
sudo dnf install calibre

calibre includes ebook-convert and supports many formats:

  • Input: EPUB, PDF, MOBI, AZW, AZW3, HTML, TXT, DOCX, ODT
  • Output: EPUB, PDF, MOBI, AZW, AZW3, DOCX, HTML, TXT

Output conventions

  • Location: Save files to research/sources/ with descriptive names
  • Naming: Use title or slug: Author-Title.pdf, title-yyyy-mm-dd.epub
  • EPUB→PDF: Check size (>100 KB typical for books)
  • PDF→EPUB: Typically 30-50% of PDF size due to reflowable format
  • Cleanup: Remove temporary files after verification

Quality checks

After EPUB → PDF:

  1. File exists and readable
  2. File size >100 KB (suspiciously small = text-only, may have lost images)
  3. file output.pdf confirms it's a valid PDF
  4. pdftoppm renders first page successfully

After PDF → EPUB:

  1. File exists and readable
  2. File size is 30-50% of original PDF (good compression ratio)
  3. Unzip the EPUB and inspect content.opf for auto-detected chapters
  4. Open in ereader or calibre viewer to verify reflowable layout

Troubleshooting

EPUB → PDF crashes or produces empty output

  • Check if EPUB is corrupt: unzip -t input.epub
  • Try without formatting options: ebook-convert input.epub output.pdf
  • If EPUB has missing stylesheets (common), calibre will warn but continue
  • Last resort: extract text from EPUB (unzip -p input.epub | grep -r "html") and convert to markdown

PDF → EPUB produces wrong chapter detection

  • Run with --verbose to see what calibre detected: ebook-convert input.pdf output.epub --verbose
  • For scanned PDFs (images), conversion may fail or produce text-only EPUB
  • Use pdftotext first to extract text, then create EPUB from markdown

File size issues

  • EPUB→PDF: if output is <50 KB, likely text-only (no images captured)
  • PDF→EPUB: if output is >150% of input, likely includes images; use --output-profile "tablet" for better compression

Notes

  • Layout preservation (EPUB→PDF): ebook-convert maintains structure, typography, images
  • Reflowable format (PDF→EPUB): EPUB reflows to device width; fixed-layout PDFs may lose structure
  • Chapter detection (PDF→EPUB): Automatic; works best on PDFs with clear page breaks or bookmarks
  • Scanned PDFs: May fail or produce text-only output (no OCR)
  • Large files: Some books convert to large files (100+ MB PDFs/EPUBs). Normal.

When conversion fails

  • EPUB is corrupt: unzip -t input.epub (EPUBs are ZIP files)
  • Verbose output: ebook-convert input.epub output.pdf --verbose for error details
  • Unsupported format: Request original source in different format
  • Manual fallback: Extract text from EPUB or PDF, save as markdown, document the failure

Related

Tested Conversions

SHELLFISH.epub → SHELLFISH.pdf (73 KB → 639 KB, layout preserved, 89 pages)
SHELLFISH.pdf → SHELLFISH-test.epub (639 KB → 262 KB, chapters auto-detected, 24 parts)


EPUB ↔ PDF bidirectional conversion pipeline ready. Last tested: 2026-04-29 | Calibre v5.x

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Research

valuemining-lengthybooks

Extract actionable insights from methodology and skill-building books via a Four-Layer framework for concept mapping, case studies, SOPs, and critical analysis.

Registry SourceRecently Updated
1370Profile unavailable
Research

biomedical-paper-billing

AI-powered biomedical manuscript generation with docx output. Activates when user provides Chinese draft/outline and requests full English research paper. In...

Registry SourceRecently Updated
1080Profile unavailable
Research

Arxiv Papers

Find and summarize arXiv.org preprints—keyword/category search, abstracts, PDF links. Use for literature scans, paper IDs, or quick orientation (not peer-rev...

Registry SourceRecently Updated
1270Profile unavailable
Research

Auto Research Pipeline

OpenClaw 原生的自动化研究 pipeline。从一个研究 topic 出发,经过 23 个 stage 产出完整论文。 每个 Phase 由独立 sub-agent 执行(context 隔离),Phase 间通过文件系统传递产出。 触发词:Research X、跑研究、文献调研、写论文、研究 pipel...

Registry SourceRecently Updated
1050Profile unavailable