Static site generators

Static sites are agentprose’s home ground: themes already emit JSON-LD, OpenGraph, a feed and a sitemap, which agentprose reads without configuration. Add the tag to the shared layout, and optionally build an index after the site build.

Hugo

Create layouts/partials/agentprose.html:

{{- /* agentprose: agent-readable content via WebMCP */ -}}
<script src="https://cdn.jsdelivr.net/npm/agentprose@0/dist/agentprose.min.js" defer
  {{- with site.Params.newsletter }} data-subscribe="{{ . }}"{{ end }}
  {{- if ne (urls.Parse site.BaseURL).Path "/" }} data-base="{{ site.BaseURL }}"{{ end }}></script>

Call it from your head partial:

{{ partial "agentprose.html" . }}

Optional: index the built site in CI.

# .github/workflows/deploy.yml (excerpt)
- run: hugo --minify
- run: npx agentprose build public --base-url "${{ vars.SITE_URL }}" --llms

Jekyll

Create _includes/agentprose.html and include it from _includes/head.html with {% include agentprose.html %}:

<script src="https://cdn.jsdelivr.net/npm/agentprose@0/dist/agentprose.min.js" defer
  {% if site.newsletter %}data-subscribe="{{ site.newsletter }}"{% endif %}
  {% if site.baseurl != '' %}data-base="{{ site.url }}{{ site.baseurl }}/"{% endif %}></script>

For a prebuilt index:

bundle exec jekyll build && npx agentprose build _site --base-url https://example.com/

Astro

In your base layout. is:inline keeps Astro from bundling the CDN script:

---
const { title } = Astro.props;
---
<html lang="en">
  <head>
    <meta charset="utf-8" />
    <title>{title}</title>
    <script is:inline src="https://cdn.jsdelivr.net/npm/agentprose@0/dist/agentprose.min.js" defer data-subscribe="/newsletter"></script>
  </head>
  <body><slot /></body>
</html>

If the site uses Pagefind (Starlight does), agentprose detects it and uses it for search_posts. Postbuild index in package.json:

{ "scripts": { "postbuild": "agentprose build dist --base-url https://example.com/" } }

Eleventy

In _includes/base.njk:

<script src="https://cdn.jsdelivr.net/npm/agentprose@0/dist/agentprose.min.js" defer></script>

In package.json:

{ "scripts": { "build": "eleventy", "postbuild": "agentprose build _site --base-url https://example.com/ --llms" } }

Plain HTML

<!doctype html>
<html lang="en">
<head>
  <meta charset="utf-8">
  <title>My post</title>
  <link rel="alternate" type="application/rss+xml" href="/feed.xml">
  <script type="application/ld+json">
  { "@context": "https://schema.org", "@type": "BlogPosting", "headline": "My post", "datePublished": "2026-10-17", "author": { "@type": "Person", "name": "Me" } }
  </script>
  <!-- The whole install: -->
  <script src="https://cdn.jsdelivr.net/npm/agentprose@0/dist/agentprose.min.js" defer data-subscribe="https://example.com/newsletter"></script>
</head>
<body>
  <article>
    <h1>My post</h1>
    <p>Agents can now read this, search the rest of the site and cite it.</p>
  </article>
</body>
</html>

Make extraction reliable

• Keep (or add) Article/BlogPosting JSON-LD or og:type=article on post pages. That’s how agentprose and agentprose build tell posts from listings.
• If the wrong element is picked as the article body, set data-content=".post-body"; strip boxes with data-ignore=".newsletter-box,.toc".
• If the blog lives under a sub-path, set data-base (the Hugo and Jekyll partials above do it for you).

Tip

Docs sites (like this one) are not “articles”. Build their index with agentprose build dist --all to include every page.

Other examples in the repository: examples/.