Anleitungen

How to Add Article JSON-LD That AI Search Engines Cite

Article JSON-LD tells AI search engines exactly what your content is, who wrote it, and why it's authoritative. This guide shows you how to implement it correctly so your pages get cited, not skipped.

Von Daniel MercerVeröffentlicht am 24. Mai 20265 Min. Lesezeit

Article JSON-LD is the single most actionable schema type for earning citations in AI-powered search engines like Perplexity, ChatGPT Search, and Google's AI Overviews. Get it right, and you give answer engines a structured, machine-readable signal that your content is authoritative, attributable, and safe to quote.

What Is Article JSON-LD?#

Article JSON-LD is a block of structured data — embedded in your page's <head> or <body> — that describes a piece of content using the Schema.org Article vocabulary. It tells crawlers the headline, author, publisher, publication date, and more in a format no parser has to guess at. Unlike microdata or RDFa, JSON-LD sits in a self-contained <script> tag, which makes it easy to inject without touching your HTML templates.

Why Do AI Search Engines Care About It?#

AI answer engines retrieve candidate passages, then rank them partly by trustworthiness signals. A page with a well-formed Article schema communicates four trust signals at once:

Authorship — author.name and author.url map to a real entity.
Recency — datePublished and dateModified tell the model how fresh the information is.
Publisher identity — publisher.name and publisher.logo anchor the content to an organization.
Canonical URL — mainEntityOfPage removes ambiguity about which URL should be cited.

Without these, the model either infers them (risking errors) or deprioritizes the page in favor of one that makes the signals explicit.

Which Article Schema Type Should You Use?#

Schema.org offers three subtypes under Article:

Article — general-purpose; safe default for most editorial content.
NewsArticle — for time-sensitive journalism; triggers Google's Top Stories carousel.
BlogPosting — for personal or brand blogs; lower authority signal than Article in most AI ranking contexts.

For how-to guides, educational content, and product explainers, use Article. Reserve NewsArticle for content with genuine news intent, because mislabeling triggers quality filters.

How to Write a Minimal Valid Article JSON-LD Block#

Paste this into your <head> and substitute your own values:

json

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "How to Add Article JSON-LD That AI Search Engines Cite",
  "description": "A practical guide to implementing Article schema markup for AI search citation.",
  "image": "https://example.com/images/article-hero.jpg",
  "author": {
    "@type": "Person",
    "name": "Jane Smith",
    "url": "https://example.com/authors/jane-smith"
  },
  "publisher": {
    "@type": "Organization",
    "name": "Example Media",
    "logo": {
      "@type": "ImageObject",
      "url": "https://example.com/logo.png"
    }
  },
  "datePublished": "2025-01-15",
  "dateModified": "2025-07-10",
  "mainEntityOfPage": {
    "@type": "WebPage",
    "@id": "https://example.com/json-ld-for-articles"
  }
}
</script>

Every field above is load-bearing. Remove dateModified and AI engines treat your content as potentially stale. Remove author.url and you lose the entity disambiguation that E-E-A-T signals depend on.

How to Extend It for Stronger AI Citation Signals#

A minimal block passes validation. A complete block earns citations. Add these properties to push further:

`wordCount`

An integer. Helps AI systems assess depth before fetching the full body.

`articleSection`

A string or array of strings matching your site's taxonomy (e.g., "SEO", "Structured Data"). This helps topic-classification models slot your page correctly.

`keywords`

A comma-separated list of terms. Not a ranking factor for traditional SEO, but useful for retrieval-augmented generation (RAG) pipelines that match queries to document metadata.

`speakable`

Points to CSS selectors or xPaths that contain the passage most suitable for voice or AI summarization. This is the closest thing to a "cite this" instruction you can give an answer engine.

json

"speakable": {
  "@type": "SpeakableSpecification",
  "cssSelector": [".article-summary", "h2", "h3"]
}

`isPartOf` and `isAccessibleForFree`

If your content sits behind a metered paywall, declare it. Undisclosed paywalls cause crawlers to index the teaser and AI engines to cite content users can't reach — a trust-killer.

How to Validate Your Article Schema#

Validation is a two-step process:

Syntax check — Paste your JSON-LD into Google's Rich Results Test. It flags malformed JSON and missing required properties.
Entity check — Run your page through the SeoChatAI Schema Generator to confirm your author and publisher entities resolve cleanly and that no fields conflict with your on-page content.

The most common error: dateModified is older than datePublished due to a copy-paste bug. Google's parser rejects the block silently.

Where Should You Place the JSON-LD Script Tag?#

Place it inside <head>. While Google's documentation says body placement is technically supported, several third-party AI crawlers — including those powering Perplexity and Bing Copilot — parse <head> first and may exit before reading body scripts on large pages. Head placement is always safer.

For CMS users:

WordPress — Use a plugin like Yoast SEO or RankMath, or add a wp_head action hook with wp_enqueue_script set to false (inline output).
Next.js — Use the <Script> component with strategy="beforeInteractive" or inject directly into <Head>.
Webflow — Paste the script block into the page's Custom Code → Head section.

What Mistakes Break AI Citability?#

Mismatched headline — headline in JSON-LD differs from the visible <h1>. AI engines cross-check these; a mismatch signals low quality.
Generic author — "author": {"@type": "Person", "name": "Admin"} provides no entity signal. Use real names with author pages.
Missing image — Several AI answer UIs display a thumbnail alongside citations. No image means no thumbnail and lower visual prominence.
Stale dateModified — Update it every time you revise the content. Treat it as a freshness timestamp, not a formality.
Multiple conflicting blocks — If your CMS and a plugin both inject Article schema, parsers pick one arbitrarily. Audit with the SeoChatAI Schema Generator to catch duplicates.

How Often Should You Update Article JSON-LD?#

Update dateModified on every substantive edit — not typo fixes, but any change that affects factual accuracy, depth, or conclusions. AI search engines weight recency differently by query type: for evergreen how-tos, a six-month-old dateModified is fine; for topics tied to software versions or regulations, anything older than 90 days risks being deprioritized in favor of fresher sources.

How to Add Article JSON-LD That AI Search Engines Cite — illustration 1

Probieren Sie unser kostenloses JSON-LD schema generator

Artikel lesen

Artikel teilenX LinkedIn

Veröffentlicht am 24. Mai 2026

Häufig gestellte Fragen

Does Article JSON-LD directly improve Google rankings?

Not directly. Article JSON-LD is a structured data signal, not a ranking factor in the traditional sense. It enables rich results eligibility and improves how AI systems understand and attribute your content, which can indirectly increase click-through rates and citation frequency.

What is the difference between Article and NewsArticle schema?

Article is a general-purpose type suited to educational, editorial, or evergreen content. NewsArticle signals time-sensitive journalism and can trigger Google's Top Stories carousel. Using NewsArticle for non-news content may trigger quality filters, so only use it when the content genuinely qualifies as news.

Do AI search engines like Perplexity and ChatGPT read JSON-LD?

Yes. AI answer engines and their underlying crawlers parse structured data to resolve authorship, publication dates, and canonical URLs. Pages with complete, valid Article JSON-LD are easier to attribute and therefore more likely to be cited in AI-generated answers.

Is the speakable property worth adding to Article schema?

Yes, particularly for AI Overviews and voice search. The speakable property points to the most citable passage on your page using CSS selectors. It functions as a direct instruction to answer engines about which text to extract, making it one of the highest-leverage additions to Article JSON-LD.

How do I fix a JSON-LD validation error in Google's Rich Results Test?

Start by checking for malformed JSON — missing commas, unclosed brackets, or incorrect quote types. Then verify all required properties are present and that field values match what's visible on-page. The Rich Results Test surfaces specific error messages; fix each one top-down before re-testing.

Can I have multiple JSON-LD blocks on one page?

You can have multiple blocks for different schema types (e.g., Article plus BreadcrumbList). However, you should never have two separate Article blocks on the same page. Duplicate Article schemas cause parsers to pick one arbitrarily, potentially discarding the more complete version.

Should the JSON-LD headline match the H1 exactly?

It should match closely, though minor formatting differences are acceptable. AI systems and Google's quality algorithms cross-reference the JSON-LD headline against the visible H1. Significant mismatches are treated as a low-quality signal and can suppress rich result eligibility.