Each chapter you want to index must use YAML frontmatter at the top of the file:

---
title: "My first post"
date: 2025-12-07
author: "Jane Doe"
description: "Short summary shown in lists."
collection: "blog"
tags:
  - rust
  - mdbook
draft: false
---

# My first post

Body content…

You can also use tags: ["rust", "mdbook"] if you prefer this syntax.

The frontmatter is mapped to this struct:

pub struct FrontMatter {
    pub title: String,
    pub date: Option<DateTime<Utc>>,
    pub author: Option<String>,
    pub description: Option<String>,
    pub collection: Option<String>,
    pub tags: Option<Vec<String>>,
    pub draft: Option<bool>,
}

Required vs Optional

• title (required)

  • If parsing fails or there is no frontmatter, title falls back to the file stem (e.g. posts/hello-world.md → "hello-world")

• date (optional but strongly recommended)

  • Accepted formats:
    • Full RFC 3339, for example 2025-12-07T10:30:00Z
    • Simple data YYYY-MM-DD, for example 2025-12-07
  • If date is missing or invalid, the file's modification time is used instead (if available)
  • Entries are sorted newest -> oldest by this value.

• author (optional)

  • Free-form string, passed through as-is into ContentEntry.author

• description (optional)

  • Short summary you want to prefer for previews
  • If missing, the body text is used instead when generating preview_html

• collection (optional)

  • Free‑form bucket name such as "blog", "notes", "changelog".

  • Used downstream to group entries (for example into collections.blog or collections.notes on the loader side).​

  • If omitted, the entry just has collection: null in JSON.

• tags (optional)

  • YAML array of strings, for example tags: ["rust", "mdbook"].

  • Missing tags become an empty [] in the JSON.​

• draft (optional)

  • Boolean flag true / false.

  • The index JSON includes draft as given; consumers like mdbook-content-loader can filter out draft: true entries before rendering.​

Behavior with missing or invalid frontmatter

If the YAML frontmatter is missing or cannot be parsed, the file is still indexed with sensible defaults:

• title: file stem.

• date: file modification time (if available), otherwise null.

• author: null.

• description: full markdown body.

• collection: null.

• tags: [].

• draft: false.​

The generated content-collections.json has one entries array, where each entry is:

{
  "path": "relative/path.md",
  "title": "...",
  "date": "2025-12-07T00:00:00Z",
  "author": "Jane Doe",
  "description": "Short summary...",
  "collection": "blog",
  "tags": ["rust", "mdbook"],
  "draft": false,
  "preview_html": "<p>First paragraphs of the body…</p>"
}

preview_html is automatically derived from the markdown body: leading boilerplate (title, TOC‑like blocks) is stripped, then up to the first 3 paragraphs / 800 characters are rendered to HTML.​

We can take advantage of this generated content-collections.json to show a rendered preview of your last modified content. This is just an example and not production ready but pretty cool nonetheless.

Extend your theme/index.hbs scroll down and find the <main> element and place the "content-collections-list" within it.

Find this code block in index.hbs:

<main>
    {{{ content }}}
    <!-- place "content-collections-list" snippet here-- >
</main>

And place the following code right below {{{ content }}}:

  <div id="content-collections-list" class="content-collections-list">
    <!-- Populated by mdbook-content-collections -->
  </div>

   <script>
    (function () {
      var indexUrl = window.location.origin + window.location.pathname.replace(/\/[^\/]*$/, '') + '/content-collections.json';

      if (window.location.protocol === 'file:') {
        indexUrl = 'content-collections.json';
      }

      fetch(indexUrl)
        .then(function (res) {
          if (!res.ok) throw new Error('Failed to load content-collections.json');
          return res.json();
        })
        .then(function (data) {
          if (!data || !Array.isArray(data.entries)) return;

          var container = document.getElementById('content-collections-list');
          if (!container) return;

          var entries = data.entries
            .filter(function (e) {
              return !e.draft && (!e.collection || e.collection === 'blog');
            })
            .slice(0, 10);

          if (entries.length === 0) {
            container.textContent = 'No posts yet.';
            return;
          }

          var list = document.createElement('ul');
          list.className = 'content-collections-items';

          entries.forEach(function (e) {
            var li = document.createElement('li');
            li.className = 'content-collections-item';

            var link = document.createElement('a');
            var htmlPath = e.path.replace(/\.md(?:own|arkdown)?$/i, '.html');
            link.href = htmlPath;
            link.textContent = e.title || e.path;

            var meta = document.createElement('div');
            meta.className = 'content-collections-meta';

            if (e.date) {
              var d = new Date(e.date);
              meta.textContent = d.toISOString().slice(0, 10);
            }

            var preview = document.createElement('div');
            preview.className = 'content-collections-preview';
            preview.innerHTML = e.preview_html || '';

            li.appendChild(link);
            if (meta.textContent) li.appendChild(meta);
            li.appendChild(preview);
            list.appendChild(li);
          });

          container.innerHTML = '';
          container.appendChild(list);
        })
        .catch(function (err) {
          console.warn('mdbook-content-collections:', err);
        });
    })();
  </script>

Now, your books landing page will have an extension above the prev-next chapter buttons showing a rendered list/overview of your last modified content. Taken from the generated content-collections.json: