Schema Graph Visualizer

← All tools

Schema Graph Visualizer

See structured data as a map: entities, properties, @id links and ghost references. Get a health score, rich-result readiness, and an editable JSON view. Fetch a page or paste raw JSON-LD.

Try it on

How this tool works

You give one URL, the server fetches the HTML itself and extracts every structured-data source on the page: JSON-LD blocks, microdata items and RDFa entities. The graph is then built and analysed in your browser, where each object becomes a node, each property a labelled edge, and every @id reference is resolved to the node that defines it.

No third-party schema service is involved, no key and no account. The page is fetched with redirects checked hop by hop against private and internal addresses, and the answer is cached for 6 hours per URL. In paste mode, the whole analysis runs locally from the JSON-LD you provide.

How the Schema Graph Visualizer works

Most schema tools give you a wall of JSON and a list of types. This one gives you the shape of the data. Enter a URL and the server fetches the page, extracts every JSON-LD block, microdata item and RDFa entity, then the browser turns that into an interactive graph: one node per entity, one labelled edge per property, and every @id reference resolved to the node that actually defines it.

Page HTMLServer parserJSON-LD · microdata · RDFaArticlePersonOrganizationauthorpublisher
Every object becomes a node, every property an edge, and references are merged onto the entity they point to.

What the graph actually shows

A node is a real-world thing the page describes: an article, a person, a company, a breadcrumb trail, an image. An edge is a property that links two things: author, publisher, isPartOf, image. Instead of reading nested brackets, you see whether the page describes a connected entity graph or a pile of isolated fragments.

Why @id is the part that matters

Serious structured data does not repeat the organization on every node. It declares it once with an @id, then other nodes point at that identifier. The graph resolves those pointers, so an article, its author, the publisher and the website collapse into one readable map.

If the page references an @id that it never defines, the tool shows a ghost node with a dashed outline and flags it. Google cannot merge what is not there, so those broken references are usually the fastest fix on the page.

The checks behind the health score

The score is a reading aid, not a Google metric. It drops for invalid JSON, undefined @id references, duplicate identifiers, and required properties missing from the types that power rich results. Every deduction is listed with the node and property involved, so you can go straight to the source.

Rich result readiness

A second panel maps the page against the property sets Google asks for on Article, Product, Breadcrumb, FAQ, Video, Event, Recipe, JobPosting, Organization and LocalBusiness. Each card says whether the type is present, partial or absent, and which required fields are missing.

One nuance worth knowing: FAQ rich results have been limited to a small set of authoritative government and health sites since 2023. The type can still be valid and useful for machines without producing an FAQ block on the SERP.

Three ways to read the same data

The graph is for understanding, the outline is for scanning and keyboard navigation, and the JSON panel is for editing. Click any node to inspect its properties, copy its object, or trace its incoming and outgoing links. Export the full payload as JSON, a PNG of the map, or a Mermaid diagram to drop into a client audit.

What it does not do

It does not pretend to be Google. It does not tell you whether a rich result will appear, it does not judge the quality of the content, and it does not fetch JavaScript-rendered schema that only exists after client-side execution. For a final verdict, run the URL through Google's Rich Results Test.

Frequently asked questions

What is a schema graph?

It is a visual representation of the entities a page declares and the properties that connect them. Nodes are objects such as Article, Person or Organization, edges are properties such as author or publisher, and @id references are merged onto the node that defines them.

Which structured data formats are supported?

JSON-LD, microdata (itemscope/itemprop) and a useful subset of RDFa. JSON-LD is the format Google recommends, and it is also the one most sites use.

Why does a node appear with a dashed outline?

The page references an @id that it never defines. Other nodes point to it, but no object in the page carries that identifier, so the entity cannot be resolved.

Is the health score an official Google score?

No. It is a local reading aid computed from syntax errors, missing required properties and broken references. Google does not publish a structured-data score.

Can I visualise a snippet without a live URL?

Yes. Switch to the paste tab and drop raw JSON-LD. Everything is analysed in your browser, with no network request.

Does it see schema injected by JavaScript?

No. The server reads the HTML as delivered. Structured data added later by a tag manager or a front-end framework will not appear unless it is also present in the initial HTML.

Is my URL stored?

The URL and the extracted payload are cached for 6 hours to avoid re-fetching the same page, then discarded. Nothing is published or shared.

No third-party schema service: the page is fetched server-side and parsed locally. JSON-LD, microdata and RDFa are read from the delivered HTML, then the graph is built in your browser. Cache: 6 hours per URL.