Audit schema markup with AI
Inspect the structured data a page actually declares, check it against visible facts and draft JSON-LD that can be validated before publication.
When you audit schema markup with AI, begin with the page visitors see. A shop may show a current price while its structured data still carries last month’s offer. The schema-markup skill distinguishes markup observed in the source or rendered page from an issue merely suggested by a crawl report.
It can then draft a correction using verified facts. An article should not acquire a rating, author or review simply because a template has room for that property. General Schema.org validity and eligibility for a Google search feature are checked separately.
What it does
- Identify declared entities in the source HTML or rendered page.
- Compare important properties with visible content and other reliable evidence.
- Choose appropriate entity types for the page’s real purpose.
- Draft JSON-LD and mark unresolved placeholders as unready for deployment.
- Separate vocabulary validation from Google rich-result requirements.
Install and use
- Install the skill and provide an authorised page, HTML or rendered output. For live Semrush Site Audit evidence, connect the official MCP and use an eligible project with a crawl snapshot.
- Ask, “Compare this product’s structured price and availability with its visible offer. Draft a correction only for the facts you can verify.”
- Validate the draft with the relevant tools and inspect the rendered page again after any separate implementation.
FAQ
Source
Made by Semrush Inc. Original skill: schema-markup, published on GitHub under the Apache-2.0 licence.
View on GitHub