Making your posts findable

Render on the server. Everything else depends on it

Google will run your JavaScript, eventually, most of the time. Almost nothing else will. The crawlers behind AI assistants, the preview bots behind Slack, WhatsApp, LinkedIn and X, and every tool that reads a page without a browser all take the HTML you serve and stop there. If your posts arrive after the page loads, those readers see an empty shell.

So fetch your posts where the HTML is made: a server component, getStaticProps, a build step, a PHP page, a template. The browser SDK entry point exists for reporting views, not for fetching the article. If you can view source on a post and not find its text, no amount of tags below will help.

What goes in the head of each post

Every post carries an seo block, and every key in it is always present, so you can map it straight into your framework’s metadata without checking whether a field exists. Each one falls back to something sensible: leave a field empty in the dashboard and use the post’s own title and excerpt.

export async function generateMetadata({ params }) {
  const { slug } = await params
  const post = await cms.posts.bySlug(slug)
  const url = `https://acme.com/blog/${post.slug}`

  return {
    title: post.seo.title || post.title,
    description: post.seo.description || post.excerpt,
    alternates: { canonical: post.seo.canonicalUrl || url },
    openGraph: {
      type: 'article',
      url,
      title: post.seo.title || post.title,
      description: post.seo.description || post.excerpt,
      images: post.seo.ogImage || post.featuredMedia?.url,
      publishedTime: post.publishedAt,
      modifiedTime: post.updatedAt,
    },
    ...(post.seo.noIndex && { robots: { index: false, follow: true } }),
  }
}

Your sitemap, without writing the loop

GET /v1/blog/sitemap lists every published post with the address it has on your site, the last time it changed, and nothing that is a draft, embargoed, or marked noIndex. The addresses are built from your project’s website and its blog path, which you set once in project settings.

# /sitemap.xml on your site proxies to:
https://api.cmskite.com/v1/blog/sitemap?format=xml&key=csk_live_…

# Or take JSON and fold it into a sitemap of your own:
curl -H "Authorization: Bearer $CMSKITE_KEY" \
  https://api.cmskite.com/v1/blog/sitemap

export default async function sitemap() {
  const { data } = await fetch(
    'https://api.cmskite.com/v1/blog/sitemap',
    { headers: { Authorization: `Bearer ${process.env.CMSKITE_KEY}` } },
  ).then((r) => r.json())

  return data
    .filter((e) => e.url)
    .map((e) => ({ url: e.url, lastModified: e.lastModified }))
}

Structured data, so an assistant can quote you properly

A block of JSON-LD tells a machine which text is the article, who wrote it and when — rather than leaving it to guess from your markup. It is what turns a result into one with a date and a byline on it, and it is what an AI assistant reads when deciding whether it can attribute an answer to you.

const ld = {
  '@context': 'https://schema.org',
  '@type': 'BlogPosting',
  headline: post.seo.title || post.title,
  description: post.seo.description || post.excerpt,
  image: post.seo.ogImage || post.featuredMedia?.url,
  datePublished: post.publishedAt,
  dateModified: post.updatedAt,
  author: { '@type': 'Person', name: post.author?.name ?? 'Acme' },
  mainEntityOfPage: `https://acme.com/blog/${post.slug}`,
}

<script
  type="application/ld+json"
  dangerouslySetInnerHTML={{ __html: JSON.stringify(ld).replace(/</g, '\\u003c') }}
/>

The small things that are already done for you

• A feed at /v1/blog/feed, in RSS, Atom or JSON, with links that point at your site rather than back at the API.
• Images compressed, resized and named after their slug, with the alt text you wrote — the only description of a picture a search engine can read.
• width and height on every image in the response, so your pages reserve the space and do not shift as they load. Layout shift is scored.
• wordCount and readingMinutes, derived once on our side so every page of yours agrees.
• Its last 20 slugs are kept, so renaming a post never costs you a link.

Five minutes, once

1. 01

   View source on a live post

   Find a sentence from the middle of the article in the HTML. If it is not there, nothing else on this page matters yet.

2. 02

   Set where posts live

   Project settings → General. /blog/{slug} or whatever yours is.

3. 03

   Map the seo block into your metadata

   Title, description, canonical, image, noIndex. The snippet above is the whole of it.

4. 04

   Point your sitemap at ours

   Then submit it once in Google Search Console.

5. 05

   Add the JSON-LD block

   And check one post in Google’s Rich Results Test.