| name | kindle-book |
| description | Convert markdown (.md) or plain text (.txt) files into Amazon KDP-ready book files: EPUB3 for Kindle digital publishing and a print-ready PDF for KDP paperback (6×9", 5×8", or custom trim). Use this skill whenever the user wants to: publish a book on Amazon Kindle, create an ebook from markdown, generate a KDP-ready EPUB or print PDF, self-publish a book, convert a manuscript to Kindle format, or make digital + print book files. Trigger even for partial phrases like "make it a Kindle book", "publish this as an ebook", "KDP files", "format my book for Amazon", or "I want to sell this on Kindle".
|
Kindle Book Skill
Converts MD/TXT manuscripts → EPUB3 (Kindle digital) + print-ready PDF (KDP paperback) using pandoc + XeLaTeX.
Quick Start
- Read the user's source file(s) from
/mnt/user-data/uploads/
- Extract or ask for book metadata (see Metadata section)
- Run
scripts/build_kindle.py with appropriate flags
- Present both output files to the user
Prerequisites Check
which pandoc xelatex python3
All three must be present. They are pre-installed in Claude's container.
Step 1: Extract Metadata
Parse YAML frontmatter from the MD file if present. If missing, ask the user for:
| Field | Required | Default |
|---|
title | ✅ | — |
author | ✅ | — |
language | ✅ | en-US |
publisher | optional | — |
rights | optional | — |
description | optional | — |
trim_size | optional | 6x9 |
cover_image | optional | none |
Supported trim sizes: 5x8, 6x9 (most popular), 8.5x11
YAML frontmatter format (tell user to add to top of MD file):
---
title: "My Book Title"
author: "Author Name"
language: en-US
publisher: "Publisher Name"
rights: "© 2025 Author Name. All rights reserved."
description: "One-line book description for metadata."
trim_size: "6x9"
cover_image: "cover.jpg"
---
Step 2: Prepare the Manuscript
Heading Structure (critical for chapter splitting)
# Title → Book title (optional, can be in frontmatter)# Chapter Name → Each # heading becomes a new chapter in EPUB## Section → Sub-sections### Sub-section → Sub-sub-sections
Markdown Features Fully Supported
- Bold
**text**, italic *text*
- Block quotes
> text
- Ordered and unordered lists
- Horizontal rules
--- (scene breaks)
- Images
- Tables
Common Issues to Fix Before Building
- Remove any existing page numbers or headers from the source
- Ensure each chapter starts with a
# heading (not ##)
- Images referenced must exist relative to the source file
Step 3: Build the Files
Run the build script:
python3 /path/to/skill/scripts/build_kindle.py \
--input /mnt/user-data/uploads/manuscript.md \
--output /mnt/user-data/outputs/ \
--title "Book Title" \
--author "Author Name" \
--language en-US \
--trim 6x9
Optional flags:
--cover /mnt/user-data/uploads/cover.jpg
--publisher "Publisher Name"
--rights "© 2025 Author"
--description "Book tagline"
--epub-only
--pdf-only
--font-size 12
The script outputs:
<slug>_digital.epub — upload to KDP eBook section<slug>_print.pdf — upload to KDP paperback interior section
Print PDF overflow prevention (built into the script):
- Long code block lines are auto-wrapped to fit the trim size
- Tables get proportional column widths via a lua filter
- Inline monospace (
code) text can break at hyphens/slashes
- URLs break at any character via xurl
- Global
\sloppy + \emergencystretch prevents text overflow
Step 4: Verify Outputs
After building, tell the user:
For digital EPUB:
- Upload to KDP → Kindle eBooks → Manuscript upload
- KDP accepts EPUB3 directly
- Preview in KDP's online previewer before publishing
For print PDF:
- Upload to KDP → Paperback → Manuscript upload
- Margins are set for the selected trim size
- Page numbers are in the header (outside edge)
- Bleed is NOT included (no full-bleed images)
- Confirm page count matches KDP's estimate before ordering proof
Troubleshooting
| Problem | Fix |
|---|
lmodern.sty not found | Script uses custom template that bypasses lmodern — use build_kindle.py, not raw pandoc |
openBinaryFile: does not exist | YAML cover-image: field is empty — remove it or set a valid path |
| Chapters not splitting in EPUB | Each chapter must start with # (H1), not ## |
| Images missing in EPUB | Image paths must be relative to manuscript file location |
| PDF text too small | Use --font-size 13 or --font-size 14 |
| KDP "text outside margins" on tables | The build script uses a lua filter to auto-assign column widths — this should be automatic. If a specific table still overflows, shorten the cell content (especially inline code) |
| KDP "text outside margins" on code blocks | The build script wraps code lines to fit the trim size. If issues persist, check for very long strings without spaces (URLs, file paths) and manually add line breaks in the markdown |
| KDP "text outside margins" on inline code | The LaTeX template allows monospace to break at hyphens. For extremely long inline code strings, break them with spaces in the markdown |
| KDP "removed non-printable markup" | Normal — KDP strips hyperlinks from print PDFs since they're not clickable on paper. No action needed |
KDP Upload Checklist (tell user)
eBook:
Paperback:
References
references/trim-sizes.md — KDP trim size margin specs for all sizesassets/print-template.tex — LaTeX template used for print PDFassets/epub-styles.css — CSS injected into EPUBscripts/build_kindle.py — Main build script
Read references/trim-sizes.md if the user requests a non-standard trim size.
