Coming from Zola

dodeca's content model is intentionally Zola-compatible. If you've used Zola, most concepts transfer directly.

What's the same

  • Content structure — content/ directory with _index.md for sections, other .md files for pages
  • TOML frontmatter — +++ delimiters, same fields like title, weight, extra
  • Template inheritance — {% extends "base.html" %}, {% block %}, {% include %}
  • Section/page distinction — _index.md creates a section, everything else is a page
  • Familiar syntax — Jinja-like templates with {{ }} and {% %}

What's different

Configuration

Zola uses config.toml. dodeca uses .config/dodeca.styx with Styx syntax:

styx
content content
output public

syntax_highlight {
    light_theme github-light
    dark_theme tokyo-night
}

Styx uses key value pairs — no colons, no equals signs. See Configuration.

Templates

dodeca uses gingembre instead of Tera. The syntax is nearly identical (Jinja-like), but there are differences in available filters, functions, and tests. See the Template Reference.

config.title and config.description

In dodeca, these come from the root _index.md frontmatter, not from the config file:

markdown
+++
title = "My Site"
description = "A site about things"
+++

Assets are processed automatically

In Zola, you opt into image resizing with resize_image(). In dodeca:

  • Images are automatically converted to JXL + WebP at multiple widths with <picture> elements
  • Fonts are automatically subsetted to only the characters used on your site
  • All assets get content-hash filenames for cache busting
  • SASS is compiled and CSS is processed automatically

Dev = Production

Zola's dev server skips some production processing. dodeca doesn't — what you see in ddc serve is exactly what ddc build produces. This means no surprises at deploy time.