Metadata and SEO

Blogin emits per-page metadata for search engines and social cards, and a site-wide robots.txt, without any template work beyond a single call in the page head.

head-meta#

A layout's base.haml head calls head-meta, which the scaffold includes:

%head
  %title= site-title
  != head-meta

For a post page it emits a canonical link and the article metadata:

<link rel="canonical" href="https://example.com/posts/hello/"/>
<meta name="description" content="A short description."/>
<meta property="og:type" content="article"/>
<meta property="og:title" content="Hello"/>
<meta property="og:description" content="A short description."/>
<meta property="og:url" content="https://example.com/posts/hello/"/>
<meta property="og:site_name" content="My Site"/>
<meta property="article:published_time" content="2026-07-20"/>
<meta property="article:modified_time" content="2026-08-02"/>
<meta property="article:tag" content="cpp"/>
<meta name="twitter:card" content="summary"/>
<meta name="twitter:site" content="@example"/>
<meta name="twitter:creator" content="@example"/>
<meta name="twitter:title" content="Hello"/>
<meta name="twitter:description" content="A short description."/>
<link rel="alternate" type="application/atom+xml" title="My Site" href="https://example.com/feed.xml"/>
<script type="application/ld+json">{"@context":"https://schema.org","@type":"BlogPosting","headline":"Hello","url":"https://example.com/posts/hello/","description":"A short description.","datePublished":"2026-07-20","dateModified":"2026-08-02","author":{"@type":"Person","name":"Pat Lee"}}</script>

The description is the post's front-matter description, falling back to the derived summary. The canonical and og:url links are absolute, built from base-url in blogin.json, so set base-url for them to be complete. A listing page emits the same tags with og:type of website, without the description tags, the article: tags, or a BlogPosting. Its title is the section label: the site title on the root listing, the term on a term page, and the humanized taxonomy name on a taxonomy index.

The article: tags appear on post pages only. article:published_time is the post's date, article:modified_time is its updated date, and there is one article:tag per tag. Each is left out when the post does not set it.

twitter:site and twitter:creator come from twitter in blogin.json, written as given, so include the @. Without it both are left out.

There is one <link rel="alternate"> per format in feed-formats, so browsers and feed readers can find the site-wide feeds from any page. A site with no posts writes no feeds and so links none.

Structured data#

head-meta writes JSON-LD that search engines read for rich results. A post page gets a BlogPosting with the headline, URL, description, dates, share image, and author from blogin.json, leaving out any of these the page does not have. The site's front page gets a WebSite. A root index.md is a post page, so it gets both:

<script type="application/ld+json">{"@context":"https://schema.org","@type":"WebSite","name":"My Site","url":"https://example.com/"}</script>

A < in a title or description is written as \u003c, so no text can end the script element.

noindex#

These pages carry <meta name="robots" content="noindex"/> and are left out of the sitemap:

  • The 404 page.
  • Every listing page after the first.
  • A post with noindex: true in its front matter.

Share images#

A link preview on LinkedIn, Facebook, Slack, or X shows an image when the page names one. Set image in a post's front matter:

---
title: Hello
image: /assets/images/hello.png
---

Set image in blogin.json for the image every other page uses, listings included:

{
  "base-url": "https://example.com",
  "image": "/assets/images/share.png"
}

With an image, head-meta adds these tags and switches the Twitter card to the large one:

<meta property="og:image" content="https://example.com/assets/images/hello.png"/>
<meta property="og:image:width" content="1200"/>
<meta property="og:image:height" content="627"/>
<meta name="twitter:card" content="summary_large_image"/>

The path is a URL on the site, so /assets/images/hello.png is the file at assets/images/hello.png and /hello.png is the file at static/hello.png. A theme's assets/ and static/ are searched after the site's own. og:image is made absolute from base-url, and with "fingerprint": true an image under assets/ is named by its hashed file. An image in static/ keeps its name. A full https:// URL is written as given.

The width and height are read from the file's header, for PNG, JPEG, GIF, and WebP. For any other format, or for an image on another host, the two tags are left out. The build warns when a named image is missing or its size cannot be read. When the image's size changes, the next build rewrites every page that uses it.

Make share images 1200 by 627 pixels. LinkedIn shows that size as a large card and a smaller image as a thumbnail. After deploying, the LinkedIn Post Inspector shows the tags LinkedIn read and refreshes its cached preview.

robots.txt#

A build with at least one post writes robots.txt at the output root, unless static/robots.txt supplies your own:

User-agent: *
Allow: /
Sitemap: https://example.com/sitemap.xml

The Sitemap line appears when base-url is set. Turn the file off with "robots": false in blogin.json.