Ebook Engine

Write in markdown. Build a branded PDF and a dual-theme HTML page. One config file.

uvx ebook-engine init
uvx ebook-engine build .

Live demo · Buy — $29 one-time · Source

---

What it does

Takes a directory of markdown (or HTML) files and produces:

• PDF — cover page, table of contents, running headers, syntax-highlighted code
• HTML — single-page web reader with a light/dark theme toggle, SEO meta tags, responsive layout

Everything is configured in one ebook.toml file. No templating language. No JavaScript framework. No frontmatter required.

Install

Requires Python 3.11+. Pick one:

# uv (recommended) — run without installing
uvx ebook-engine --help

# uv — install as a persistent tool
uv tool install ebook-engine

# pip
pip install ebook-engine

For PDF output, install weasyprint — the engine finds it on PATH:

brew install weasyprint   # macOS
apt install weasyprint    # Debian/Ubuntu

Quick start

mkdir my-book && cd my-book
ebook-engine init

echo "# Chapter 1
This is the first chapter." > 01-intro.md

ebook-engine build .
# → dist/my-book.pdf   (cover + TOC + chapters)
# → dist/index.html    (dual-theme web page)

The build auto-detects the source type: .md files are treated as markdown chapters, .html files as HTML intake (exported from Notion, Affine, and similar tools — export artifacts like duplicate titles and workspace backlinks are stripped automatically).

Configuration

ebook-engine init writes this ebook.toml:

[book]
title = "My Book"
subtitle = "A practical guide"
author = "Author Name"
description = "A description for meta tags and SEO."
keywords = ["topic", "guide"]

[brand]
accent = "#00E5FF"   # used for links, headings, borders
font = '"Space Mono", ui-monospace, Menlo, monospace'
footer = "grid[lab]" # shown at the bottom of every page

[output]
directory = "dist"
pdf = true           # requires weasyprint
html = true          # dual-theme light/dark page

[analytics]
vercel = false       # inject Vercel Analytics (works on Vercel deploys)
google = ""          # Google Analytics measurement ID (e.g. "G-XXXXXXXXXX")
matomo_url = ""
matomo_site_id = ""

Features

• Dual-theme HTML — light reader mode and dark branded mode, persisted in localStorage
• Configurable branding — accent color, font stack, footer text, all in ebook.toml
• SEO-ready — canonical URL, robots meta, description, keywords
• Analytics injection — Vercel, Google Analytics, or Matomo
• PDF with print CSS — running headers, page numbers, cover page, table of contents
• HTML intake cleaner — strips export artifacts from Notion, Affine, and similar tools
• Markdown with Obsidian support — wikilinks, image embeds, callouts, YAML frontmatter
• Safe by default — config values are HTML-escaped; the output directory is validated against path traversal

Not a CMS

This is a build tool, not a content management system. You write in files. You run a command. You get output artifacts. There's no database, no admin panel, no login. The source of truth is your git repo.

If you want a WYSIWYG editor with real-time collaboration, use Notion or Google Docs — then export to HTML and feed it to the engine.

Docs

Full guides live in docs/: getting started, configuration, CLI reference, examples.

License

MIT. Source: github.com/IsaacBell/ebook-engine. The $29 purchase supports development and gives you priority support. The engine is fully functional without a license key.