Structured data for ChatGPT is machine-readable information used either to describe a webpage or to control the format of an API response. There are 2 common meanings, and a sample article schema can include a publication date such as 2026-09-20:

  1. Schema.org markup on a website, usually written in JSON-LD, that explains what a page contains.
  2. OpenAI Structured Outputs, an API feature that makes ChatGPT return data in a defined JSON Schema format.

These technologies solve different problems. Schema markup helps systems interpret web pages. OpenAI Structured Outputs controls the format of model responses.

What Does Structured Data Mean for a Website?

For SEO and ChatGPT Search, structured data is machine-readable information added to a webpage. It identifies entities and attributes such as:

  • An article's headline and author
  • A product's name, price and availability
  • A business's address and opening hours
  • A recipe's ingredients and cooking time
  • A page's position in a website hierarchy

Most website structured data uses the Schema.org vocabulary and the JSON-LD format. Google describes structured data as a way to provide explicit clues about the meaning of page content. Google generally recommends JSON-LD because it is easier to implement and maintain.

A simplified article example looks like this:

<script type="application/ld+json">
{
  "@context": ""@type": "Article",
  "headline": "What Is Structured Data for ChatGPT?",
  "author": {
    "@type": "Organization",
    "name": "Example Company"
  },
  "datePublished": "2026-09-20",
  "about": {
    "@type": "Thing",
    "name": "Structured data for ChatGPT"
  }
}
</script>

This markup does not normally change what visitors see. It adds labels that help search engines and other automated systems interpret the page.

Does Schema.org Markup Make ChatGPT Cite a Website?

No. Schema.org markup does not guarantee that ChatGPT will rank, mention or cite a website.

Schema markup can clarify the entities and facts on a page, but ChatGPT Search also depends on factors such as:

  • Whether the page can be crawled
  • Whether the content matches the user's query
  • Whether the page provides trustworthy and useful information
  • Whether the information is clear and consistent
  • Whether OpenAI's systems select the page as a relevant source

OpenAI states that publishers should avoid blocking OAI-SearchBot if they want their public content to be discoverable and potentially included in ChatGPT summaries and snippets. OpenAI also states that its crawlers follow robots.txt rules.

Schema markup is a supporting signal, not a special ranking switch for ChatGPT.

What Structured Data Should a Website Use for ChatGPT?

Use the Schema.org type that accurately describes the page.

Page type Relevant structured data
Blog post or editorial article Article
News story NewsArticle
Product page Product
Local business page LocalBusiness
Organisation page Organization
Person profile Person
How-to guide HowTo
Event page Event
Site navigation BreadcrumbList
Website search page WebSite with SearchAction, where applicable

The markup should describe information that is present and accurate on the page. Google advises against adding structured data for content that is hidden, blank, misleading or unrelated to the visible page.

On a product page, structured data can identify the product name, brand, price, currency, availability and review information. Google can use valid product markup for richer search appearances, although enhanced results are not guaranteed.

What Is OpenAI Structured Outputs?

OpenAI Structured Outputs is an API feature that makes a model response conform to a developer-provided JSON Schema.

This is different from website Schema.org markup. Developers can use Structured Outputs when an application needs predictable machine-readable responses, such as:

  • Extracting customer details from an email
  • Returning product recommendations in a fixed format
  • Producing data for a user interface
  • Identifying action items in meeting notes
  • Connecting a model to software tools or databases

OpenAI provides Structured Outputs through JSON Schema response formats and function calling. When strict schema adherence is enabled, the response is designed to match the supplied schema rather than merely being valid JSON.

A conceptual schema might look like this:

{
  "type": "object",
  "properties": {
    "answer": {
      "type": "string"
    },
    "confidence": {
      "type": "number"
    },
    "sources": {
      "type": "array",
      "items": {
        "type": "string"
      }
    }
  },
  "required": ["answer", "confidence", "sources"],
  "additionalProperties": false
}

The schema tells the application which fields to expect. It does not add SEO markup to a webpage.

Schema Markup Versus OpenAI Structured Outputs

Feature Website structured data OpenAI Structured Outputs
Main purpose Explain webpage content to machines Control the format of model responses
Typical format Schema.org in JSON-LD JSON Schema
Used by Search engines and automated systems Developers using the OpenAI API
Main subject Pages, products, organisations and events Model-generated answers and tool arguments
Improves response reliability? No Yes, when the schema is supported and correctly configured
Guarantees ChatGPT visibility? No No
Main implementation location Website HTML API request

How to Make a Website Easier for ChatGPT to Understand

For a website that wants to be discoverable in ChatGPT Search, use this implementation order:

  1. Publish clear, useful page content. State the main answer, entities, dates, prices and qualifications explicitly.
  2. Use descriptive headings and semantic HTML. Make relationships between topics easy to identify.
  3. Add accurate Schema.org JSON-LD. Use Article, Product, Organization, Person or another type that matches the page.
  4. Keep structured data consistent with visible content. Update prices, availability, authorship and dates when the page changes.
  5. Allow appropriate crawling. Check that robots.txt and security tools do not block OAI-SearchBot if ChatGPT discovery is a priority.
  6. Validate the markup. Google provides the Rich Results Test and documentation for testing supported structured data types.

What Structured Data Cannot Do

Structured data cannot:

  • Force ChatGPT to cite a page
  • Replace original, accurate content
  • Make unsupported claims trustworthy
  • Hide important information from users while exposing it to AI systems
  • Guarantee a Google rich result
  • Turn a poorly structured website into an authoritative source

Google says structured data can make pages eligible for richer search features, but the search system decides whether those features appear.

Bottom Line

Use Schema.org JSON-LD when you are describing a webpage to search engines and other automated systems. Use OpenAI Structured Outputs when an application needs model responses in a defined JSON format.

Neither one guarantees that ChatGPT will include, rank or cite a website. For ChatGPT Search, combine clear visible content, accurate entity information, crawlable pages and structured data that matches what users can see.