Markdown Baseline#

Copy as MarkdownView Markdown

Common GitHub-flavored Markdown behavior is enabled by default. You do not need any configuration for the features on this page — every rendered example below is produced by this documentation site itself with the default settings.

Every option above is an extension on top of CommonMark, and each is opt-out. The parser underneath targets full conformance: it renders all 652 CommonMark 0.31.2 spec examples correctly in its core profile, checked on every CI run. A document that uses none of the extensions conforms to the specification under the conformance suite's HTML normalization rule; the markup is not byte-identical, because ox-content adds slug id attributes to headings. See CommonMark Conformance for the per-profile numbers.

Turn any of them off explicitly when a site needs stricter CommonMark behavior:

import { oxContent } from "@ox-content/vite-plugin";

export default {
  plugins: [
    oxContent({
      strikethrough: false,
      taskLists: false,
    }),
  ],
};

Tables#

| Feature    | Status  |
| ---------- | ------- |
| Tables     | Default |
| Task lists | Default |

Rendered:

Task Lists#

- [x] Parse Markdown in Rust
- [x] Render HTML
- [ ] Take over the world

Rendered:

• Parse Markdown in Rust
• Render HTML
• Take over the world

Strikethrough#

Ox Content is ~~slow~~ fast.

Rendered:

Ox Content is slow fast.

Autolinks#

Bare URLs become links. The default follows gfm, so autolinks: false opts out without giving up the rest of GFM.

Docs live at https://ubugeeei-prod.github.io/ox-content/

Rendered:

Docs live at https://ubugeeei-prod.github.io/ox-content/

Auto-linked URLs open in a new tab with rel="noopener noreferrer".

A URL ends at whitespace, and trailing punctuation is trimmed. CJK sentence punctuation ends it as well, because Japanese and Chinese prose leave no space between a URL and the mark that follows it:

詳細は https://example.com/foo。次の文。

Rendered:

詳細は https://example.com/foo。次の文。

Non-ASCII characters inside the URL itself are kept, so an IRI such as https://ja.wikipedia.org/wiki/日本語 still links in full.

Footnotes#

Ox Content renders footnotes natively.[^1]

[^1]: This is the footnote body.

Rendered:

Ox Content renders footnotes natively.¹

The reference becomes a superscript link. With the default renderer the visible marker is the source identifier, and each definition is emitted in place as <div class="footnote"> — put definitions at the bottom of a page to collect them there.

Set semanticFootnotes: true to assign stable numeric markers in document order ([^deployment-note] → 1, 2, …) and collect definitions into one accessible section. The source identifier is used only for lookup and slug generation (fn-… / fnref-…). Repeated references keep unique ids (fnref-note, fnref-note-2, …) and each occurrence gets a backlink. Definition bodies keep their block content. No client JavaScript is required.

<section class="footnotes" aria-label="Footnotes">
  <ol>
    <li id="fn-deployment-note">
      … <a href="#fnref-deployment-note" aria-label="Back to reference 1">↩</a>
    </li>
  </ol>
</section>

oxContent({
  footnotes: true,
  semanticFootnotes: true,
});

This documentation site enables semanticFootnotes so the live example above uses the ordered section. The option stays off by default so current alpha HTML does not change.

Frontmatter#

YAML frontmatter is parsed before rendering and never appears in the output HTML. This page starts with:

---
title: Markdown Baseline
description: GitHub-flavored Markdown, frontmatter, and table-of-contents defaults that are on out of the box.
---

The SSG theme uses title for the document title and navigation, and description for <meta name="description"> and Open Graph tags. Any other keys are passed through: .md modules expose them as the frontmatter export, collections expose them to queries, and custom transformers receive them as context.frontmatter.

import { frontmatter, html } from "./guide.md";

console.log(frontmatter.title); // "Markdown Baseline"

Table of Contents#

A TOC is built from headings during transform. The sidebar navigation for this very page is driven by it. tocMaxDepth: 3 means # through ### are included by default; deeper headings are rendered but not indexed.

oxContent({
  toc: true,
  tocMaxDepth: 3,
});

The TOC is exposed to .md modules as a tree of { depth, text, slug, children } entries:

[
  {
    "depth": 1,
    "text": "Install Guide",
    "slug": "install-guide",
    "children": [
      { "depth": 2, "text": "Prerequisites", "slug": "prerequisites", "children": [] },
      { "depth": 2, "text": "Run Vite", "slug": "run-vite", "children": [] }
    ]
  }
]

Every heading also gets a stable id attribute (the slug above), so deep links like #task-lists work on every page.

Visible # permalink controls next to the heading are opt-in. See Heading Permalinks. Off by default so existing HTML stays unchanged.

[[toc]] in the body#

A paragraph holding nothing but [[toc]] becomes the page outline, in place. Useful on a long page a reader arrives at from a deep link, and on narrow viewports where the right-hand outline (ssg.theme.aside) is hidden:

# Glossary

[[toc]]

## Contents

The casing does not matter — [[TOC]] and [[Toc]] work the same way — and tocMaxDepth decides how deep the list goes. Anything else on the line leaves the text literal, so a page documenting the directive can still show it, and so can a fenced code block.

The outline is part of the generated HTML, so it is in the static page and in the search payload. With wiki links enabled, [[toc]] stays the directive; link a page actually named toc with [[toc|toc]].

Related#

• Heading Permalinks — opt-in visible # links on those ids.
• Syntax Extensions — opt-in authoring syntax on top of this baseline.
• NotByAI Badge — opt-in static human-authorship disclosure.
• Built-in Features overview