Pre-release

This release is a pre-release and may not be stable for production use.

Jinjabread

A Python-based static site generator using Jinja templates.

Inspired by staticjinja, jekyll, and hugo.

Install

Requires Python 3.11 or newer.

pip install jinjabread

Usage

Create new site project

python -m jinjabread new mysite

Build site

python -m jinjabread build mysite

Preview site locally

python -m jinjabread serve mysite
# Visit http://127.0.0.1:8000 in your browser.

Features

Write pages in Markdown, HTML, or text.
Use Jinja2 templating language in Markdown, HTML, or text.
Supports YAML metadata in Markdown pages.
Keep static media alongside static pages.
Index pages (i.e., index.html) can list all other pages in the same directory.
Preview your static site locally with a built-in web server.
Prettify all generated HTML (because why not?)

HTML pretty-printing

With prettify_html enabled (the default), jinjabread re-indents every generated page for readability. It is a rendering-preserving reformatter rather than a byte-preserving one, and it makes a specific contract:

It preserves how the page renders. The output displays identically to the input: inline elements are never broken across lines, significant whitespace is kept, and pre, textarea, script, and style content is emitted verbatim. Only insignificant whitespace between block-level elements changes.
It is idempotent. Prettifying already-pretty output changes nothing.
It ends every file with a single newline. Non-empty output is terminated with exactly one \n, following the POSIX convention that a text file's last line ends in a newline. Empty input stays empty.
It adds no wrapper of its own. A body-level fragment, such as several top-level <p> elements or plain text, is emitted as-is and is never wrapped in a <div> or <span>.
It parses and repairs markup. Every page round-trips through lxml's HTML parser, so unclosed tags are closed and misnested tags are corrected, matching how a browser reads the input. The output is always well-formed HTML; invalid or non-standard structure is not preserved verbatim.
It distinguishes documents from fragments. A whole document, or a document-level element such as a lone <script> or <head>, is emitted as a complete document with an HTML5 doctype; a body-level fragment is emitted without a doctype or wrapper.

File structure

Important files and directories

Name	Description	Example path
Project directory	The project root	`mysite`
Content directory	Contains site content where each file becomes a site page	`mysite/content`
Layouts directory	Contains page layouts that gets used by the site content	`mysite/layouts`
Static directory	Contains static media that gets copied to the output directory	`mysite/static`
Output directory	The complete generated site, ready to be hosted	`mysite/public`
Config file	Custom site configurations in TOML format	`mysite/jinjabread.toml`

Example: Site project structure

mysite/
├── content
│   ├── about.html
│   ├── index.html
│   └── posts
│       ├── index.html
│       ├── my-journey
│       │   ├── index.html
│       │   └── profile-photo.jpg
│       └── my-story.html
├── jinjabread.toml
├── layouts
│   └── markdown.html
├── public
│   ├── about.html
│   ├── index.html
│   ├── posts
│   │   ├── index.html
│   │   ├── my-journey
│   │   │   ├── index.html
│   │   │   └── profile-photo.jpg
│   │   └── my-story.html
│   └── static
│       └── style.css
└── static
    └── style.css

Example: File to URL translation

File path	URL path
`public/index.html`	`/`
`public/about.html`	`/about`
`public/posts/index.html`	`/posts/`
`public/posts/my-story.html`	`/posts/my-story`
`public/posts/my-journey/index.html`	`/posts/my-journey/`
`public/posts/my-journey/profile-photo.jpg`	`/posts/my-journey/profile-photo.jpg`
`public/static/style.css`	`/static/style.css`

Site config

Default config

content_dir = "content"
layouts_dir = "layouts"
static_dir = "static"
output_dir = "public"
prettify_html = true

[context]

[[pages]]
  type = "jinjabread.MarkdownPage"
  glob_pattern = "**/*.md"
  layout_name = "markdown.html"

[[pages]]
  type = "jinjabread.Page"
  glob_pattern = "**/*"

Custom config

Change output directory

# jinjabread.toml
output_dir = "dist"

Add global Jinja context variables

# jinjabread.toml
[context]
  site_name = "My site"
  url_origin = "https://mysite.com"

Change Markdown layout file

# jinjabread.toml
[[pages]]
  type = "jinjabread.MarkdownPage"
  glob_pattern = "**/*.md"
  layout_name = "post.html"

[[pages]]
  type = "jinjabread.Page"
  glob_pattern = "**/*"

Add page-specific Jinja context variables

# jinjabread.toml
[[pages]]
  type = "jinjabread.MarkdownPage"
  glob_pattern = "**/*.md"
  layout_name = "markdown.html"

[[pages]]
  type = "jinjabread.Page"
  glob_pattern = "**/*.txt"

  [pages.context]
    foo = "bar"

[[pages]]
  type = "jinjabread.Page"
  glob_pattern = "**/*"

Pages types

Page type	Keyword arguments
`jinjabread.Page`	- `glob_pattern`
`jinjabread.MarkdownPage`	- `glob_pattern` - `layout_name`

Markdown pages

Markdown content supports full YAML metadata.

For example, given the following content and layout:

---
# content/my-blog-post.md
title: My blog post
author: John Smith
description: A very nice story.
keywords:
  - thrilling
  - must-read
---
It was a cold stormy night...

<!-- layouts/markdown.html -->
<h1>{{ title }}</h1>
<h2>{{ description }}</h2>
<h3>Written by {{ author }}</h3>
{{ content }}

Results in the following output:

<!-- public/my-blog-post.html -->
<h1>
  My blog post
</h1>
<h2>
  A very nice story.
</h2>
<h3>
  Written by John Smith
</h3>
<p>
  It was a cold stormy night...
</p>

Context variables

Default variables

All pages have these context variables:

Name	Description	Example value
`url_path`	The URL path of the current page	`/`
`file_path`	The file path of the current page, relative to the output directory	`index.html`

Variable precedence

Page-specific context variables override global site context variables.

For example:

# jinjabread.toml
[context]
  fee = "fie"
  foe = "fum"

[[pages]]
  type = "jinjabread.Page"
  glob_pattern = "**/*.txt"

  [pages.context]
    foe = "foo"

[[pages]]
  type = "jinjabread.Page"
  glob_pattern = "**/*.html"

  [pages.context]
    foe = "bar"

.txt pages will have the following extra context variables:
```
fee = "fie"
foe = "foo"
```
.html pages have the following extra context variables:
```
fee = "fie"
foe = "bar"
```

Index pages

All index pages (i.e., index.*) has an extra context variable named pages which is list of dictionaries of context variables from its sibling files.

For example, given the following file structure:

mysite/content/posts/
├── index.html
├── post1.md
└── post2.md

You can list all the pages in the posts directory using:

<!-- mysite/content/posts/index.html -->
{% for page in pages %}
<a href="{{ page.url_path }}">
  {{ page.url_path.split('/') | last | title }}
</a>
{% endfor %}

Resulting in:

<!-- mysite/public/posts/index.html -->
<a href="/posts/post1">Post1</a> <a href="/posts/post2">Post2</a>

Contributing

Setup

Install uv (https://docs.astral.sh/uv/getting-started/installation/), then:

uv sync

Test

uv run python -m unittest

Build

uv build

Release

export TWINE_USERNAME='__token__'
export TWINE_PASSWORD='secret-token'
uv run twine upload dist/*

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

jinjabread-0.7.0a2.tar.gz (106.5 kB view details)

Uploaded Jul 23, 2026 Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

jinjabread-0.7.0a2-py3-none-any.whl (15.9 kB view details)

Uploaded Jul 23, 2026 Python 3

File details

Details for the file jinjabread-0.7.0a2.tar.gz.

File metadata

Download URL: jinjabread-0.7.0a2.tar.gz
Upload date: Jul 23, 2026
Size: 106.5 kB
Tags: Source
Uploaded using Trusted Publishing? No
Uploaded via: twine/6.2.0 CPython/3.11.15

File hashes

Hashes for jinjabread-0.7.0a2.tar.gz
Algorithm	Hash digest
SHA256	`eb4fa2f076c632aa2ae5bbef52a3a078d035532deb243982b23034a5f188ca86`
MD5	`6606b28485e31a589ef23a99fed02810`
BLAKE2b-256	`54aa3e966dc89c63f12a37e3bbe20c0784547c47560479375e78d39cb1598aba`

See more details on using hashes here.

File details

Details for the file jinjabread-0.7.0a2-py3-none-any.whl.

File metadata

Download URL: jinjabread-0.7.0a2-py3-none-any.whl
Upload date: Jul 23, 2026
Size: 15.9 kB
Tags: Python 3
Uploaded using Trusted Publishing? No
Uploaded via: twine/6.2.0 CPython/3.11.15

File hashes

Hashes for jinjabread-0.7.0a2-py3-none-any.whl
Algorithm	Hash digest
SHA256	`9198b7b938ff13548999c72644d3783468f11b08c071e51f8a8fe103e8c1dea6`
MD5	`d7ddf27ffb6676849c0870e157e262ea`
BLAKE2b-256	`737ce8fbf4f0a4f21357ad359dd9e3580e244a9aa5416632e99f6689b2022286`