Structure
FAQPage contains mainEntity, an array of Question nodes. Each Question has name and acceptedAnswer with @type Answer and text.
Data source
Use only questions and answers that visitors can actually see. Do not invent FAQ content from keywords merely to fill a schema.
{
"@context": "https://schema.org",
"@type": "FAQPage",
"name": "FAQPage example"
}Primary page type
FAQ does not replace the primary content type. A product page stays Product, an article stays Article and a service page stays Service.
HTML and JSON-LD
Questions may be extracted from accordions or FAQ components, but the parser should reject hidden JavaScript, empty containers and unrelated navigation text.
Quality
A large number of questions does not guarantee better structured data. Correct text and consistency with the rendered page matter more.
When not to generate
If only one incidental question is found or the block is not a real FAQ, it is safer not to create FAQPage automatically.
Common questions
Can structured data always be improved to 100% automatically?
Only when the required data actually exists in the JSON-LD or on the page. A tool should not invent prices, authors, ratings, addresses or other values.
Should every possible Schema.org property be added?
No. A smaller set of accurate properties is better than irrelevant or fabricated values.