Schema Builder
Auto-generate JSON-LD structured data markup for Organization, Product, FAQPage, Article, HowTo, and more from your page content and brand facts.
Overview
The Schema Builder analyzes your page content and brand facts to generate JSON-LD structured data markup. Structured data helps AI models and search engines understand your content's meaning, increasing the likelihood of being cited accurately in AI-generated answers.
The builder produces ready-to-paste <script type="application/ld+json"> blocks for multiple Schema.org types, with each schema accompanied by a confidence score explaining why it was generated.
Before generating anything, the builder detects the schema already on your page — including types nested inside @graph wrappers (Yoast, Rank Math) and top-level JSON-LD arrays — and reports the count and type list. Only missing types are generated; existing markup is never duplicated.
Supported Schema Types
| Schema Type | When Generated | Confidence |
|---|---|---|
| WebPage | When the page's title can be read and the page doesn't already carry WebPage schema. If the title can't be read, the builder reports it honestly instead of emitting placeholder markup. | 0.9 |
| Organization | When brand facts include company name, description, or founding information. | 0.85 |
| FAQPage | When the page has two or more headings whose answer text can be extracted from the body beneath them (question-format headings preferred). | 0.55--0.8 |
| Article | When page body text exceeds 300 words. | 0.6--0.8 |
| HowTo | When page headings follow step-format patterns (e.g., "Step 1:", "1."). | 0.65 |
| Product | When brand facts include product-category entries (name, price, description). | 0.6 |
How It Works
- Input — You provide a URL. The builder reuses page content stored by a prior Content Optimization run when available, and otherwise reads the page live (title, meta description, headings, body text, existing JSON-LD). It also loads your project's brand facts.
- Existing-schema detection — The page's current JSON-LD is flattened (including
@graphwrappers and array blocks) into a typed item list, shown as "Already on this page". Types that already exist are excluded from generation. - Analysis — The system evaluates the content against pattern rules for each missing schema type. For example, it counts question-format headings for FAQPage detection and measures word count for Article detection.
- Generation — Missing schema types are generated with all available fields populated. Brand facts enrich Organization and Product schemas with details like founding date, logo URL, contact email, and pricing.
- Output — All generated schemas are combined into a single JSON-LD block ready for embedding. If the page can't be fetched, the builder returns a clear error rather than markup built from placeholders.
Schema Fields
Organization
Generated from brand facts. Supported fields:
name— Fromcompany_name,brand_name, ornamebrand facts.url— Fromwebsiteorurlbrand facts.description— Fromdescriptionortaglinebrand facts.logo— Fromlogoorlogo_urlbrand facts.foundingDate— Fromfoundedorfounded_yearbrand facts.email— Fromemailorcontact_emailbrand facts.
Product
Generated when brand facts have a product category with at least two entries:
name— Product name.description— Product description.offers— Price and currency (extracted frompricebrand fact).
FAQPage
Generated when the page has two or more headings with real answer text beneath them. Each heading becomes a Question and the prose that follows it (up to the next heading) becomes the acceptedAnswer — extracted in document order from the page, not a placeholder. Question-format headings (what, who, why, how, etc.) score highest (0.8); plain headings with extractable answers still qualify at lower confidence (0.55). If no real answers can be extracted, no FAQPage is generated — the builder never emits placeholder answer text that would fail validation.
Article
Generated when the page body exceeds 300 words. Includes headline, description, URL, and word count. Higher word counts produce higher confidence scores.
HowTo
Generated when headings follow step patterns like "Step 1:" or "1. Do something". Each step heading becomes a HowToStep.
How to Use
- Navigate to Schema Builder in the sidebar.
- Enter the URL you want to generate schema markup for, or select from your project's tracked pages.
- Review the generated schemas. Each one shows the schema type, confidence score, and the reason it was generated.
- Copy the combined JSON-LD output.
- Paste it into a
<script type="application/ld+json">tag in your page's<head>section.
Tips for Better Schema Output
- Add comprehensive brand facts — The more facts you provide (company name, description, founding year, logo URL, pricing), the richer your Organization and Product schemas will be.
- Structure your page content with question-format headings to trigger FAQPage schema generation.
- Use numbered step headings for instructional content to trigger HowTo schema.
- Write long-form content (800+ words) for the highest Article schema confidence.
- Validate the generated JSON-LD using Google's Rich Results Test before deploying.
Plan Requirements
The Schema Builder requires the Advanced AI Insights feature, available on Pro-Individual plans and above.