Tool · Entity visualization
Entity Observatory
See how your site describes a person, a company and the links between them. Entity Observatory is entity visualization for structured data and Wikidata. The map shows who wrote a book, who founded a company and where the profiles lead.
Enter a page URL or paste JSON-LD. Click a relation to see the code that declares it. You can add a Wikidata record next to it as a separate layer.
The tool checks the data in the code of one public page. It does not crawl the whole domain or run JavaScript. Limit: 3 MB.
Up to 20 MB in total. File contents stay in your browser’s memory. Scripts in HTML are never run.
A separate connection to the Wikimedia API that you start yourself. Only IDs are sent, not the content of your files. Wikimedia sees your IP address. We fetch the statements of the chosen record, not all incoming relations.
Advanced options
Entity visualization and relations
| Subject | Relation type → | Object / value | Source |
|---|
Scope, sources and diagnostics
What it’s for
See whether the person, the company and the website each have their own @id or merge into one node.
Check who is the author, who is the founder and who works where. The arrow runs from the subject to the object.
Click a relation to see which block and which place in the code declares it.
Add the Wikidata record as a separate layer and compare what the page says with what the record says.
Who I am
Why I built Entity Observatory
I’m Ewelina Podrez-Siama. I’ve worked in SEO since 2009 and on brand visibility in AI since 2024. I work as an independent consultant and trainer, and I’ve written or co-written four books on SEO.
For years I’ve been explaining what an entity is. Why schema.org, when search engines read the text anyway. Why a Wikidata record, when the site already has an “About me” page. In a workshop everyone nods, and then the question comes anyway: “OK, but what does it look like on my site?”
For most people an entity is an abstract idea. JSON-LD in the page code doesn’t help, because to most of us it’s a wall of brackets. Yet inside are simple sentences: this person founded this company, this book has this author, this LinkedIn profile is the same person.
That’s why I built Entity Observatory. It turns those sentences into a map you can look at, click through and show your boss or a client. It adds nothing of its own. If something isn’t in the data, it isn’t on the map either. It’s most useful on your own data, because you see right away what the page really declares and what you only assume.
More about me on the About me page.
In short
What is Entity Observatory?
Entity Observatory is a free entity visualization tool. It shows which people, companies, books and other objects a page describes in schema.org structured data and in Wikidata, and how they are connected. Every relation is a knowledge triple: subject → relation type → object, with a link to the place in the code that declares it.
It is not an ERD. An entity-relationship diagram from database design shares the words, but answers a different question:
| Entity-relationship diagram (ERD) | Entity Observatory | |
|---|---|---|
| What it is for | designing a database | checking what a page declares about itself |
| What it shows | entity types (e.g. Customer, Order), their fields and relationship cardinality | specific entities, e.g. a person, a company, a book, and the relations between them |
| Where the data comes from | a designer draws it before the data exists | the page’s structured data and Wikidata |
What the tool does not check, e.g. whether the code is valid or what is in Google’s Knowledge Graph, is covered in What the tool doesn’t do.
How to read the map
The arrow always runs from the subject to the object
Every relation is a triple: subject → relation type → object (the jargon word is predicate). When a book has an author, the arrow runs from the book to the person, because that is how the data says it.
The short version: click an object → pick a relation → check the source.
Circle – an entity described in the data: it has its own @id or is described in place without one, e.g. a company address.
Capsule – a text value, number or date, written inside. It gets no identifier and does not pretend to be an entity. Profile URLs (Facebook, LinkedIn, Wikipedia…) carry the service badge.
Diamond – a reference to an address this import does not describe. No description does not mean the entity does not exist.
More about how the map works
Clicking an object shows its details, and a second click or the “Set as map center” button puts it in the middle. The map shows the neighborhood of one node. Every relation type gets a place on it before any type gets a second one. The number next to a label, e.g. “knowsAbout · 20”, tells you how many relations of that type there are, and the “+N” button next to it shows them on the map. Types that did not fit are under the link above the map and in the node panel.
On a phone you first get one readable triple with arrows to the next relation, and the “Show the map” button opens the whole map.
Four inputs
Where the tool gets its data
- Page URL: the podrez.pl server fetches one public address and reads the JSON-LD blocks from its HTML. No login, no JavaScript and no crawling of the whole domain. Microdata and RDFa are not read.
- File or pasted code: JSON-LD, HTML with schema or a saved Wikidata export. The analysis runs in your browser.
- Wikidata: your browser fetches the chosen record straight from the Wikimedia API, with qualifiers, references, rank and revision number. These are the record’s own statements, not every relation that points to it.
Technical details
The schema.org context is not fetched from the web. The tool has a built-in profile matching the official context as of November 2, 2026, checked against all 3,103 terms. You can add other external contexts in a local context pack.
One import holds up to 150,000 relations. The map shows a slice around the chosen node, while the table and search cover everything.
Privacy
Files stay in your browser
Your browser analyzes the file and the pasted code. Nothing goes to the server, and the data disappears when you refresh the page, unless you save a project.
URL mode works differently: the address goes to the podrez.pl server, which fetches the page and sends it back to your browser. I don’t store the address or the content. The server has request limits and does not fetch private or local addresses or unusual ports.
With Wikidata, Wikimedia sees your IP address and the ID of the record you ask for. The tool doesn’t send what you analyze to Google Analytics.
Result
What the result shows
-
Relation map – the neighborhood of the chosen node: schema, Wikidata or both layers at once, with the direction of every relation.
-
Triple table – the whole import with a filter, search and a distinction between text, number, date and identifier.
-
Source of every statement – the block and the JSON path a relation comes from. For Wikidata also qualifiers, references, rank and revision.
-
Export – a CSV of triples for a spreadsheet, the map as SVG and a project file you can reopen without sending any data.
Honest about limits
What the tool doesn’t do
- It doesn’t show Google’s Knowledge Graph. It shows what the data declares. What Google took from it is another question.
- It doesn’t measure “entity strength”. The size and position of a node on the map say nothing about authority.
- It doesn’t replace the Rich Results Test. To check whether a page qualifies for rich results, use Google’s tool.
- It isn’t a full JSON-LD 1.1 processor. Rare constructs, e.g. type-scoped contexts, reject the whole block with a message instead of giving a result that only looks right.
Tools and lists
My other tools and lists
Free, no sign-up. Each one answers one specific question.
Tool
Are you in Google’s Knowledge Graph?
Enter a name and check whether Google knows you as an entity and whether it matches Wikidata.
Tool
Is your site in Common Crawl?
See how many of your pages made it into the Common Crawl archive and whether anything blocks CCBot.
Tool
AI bots in logs
Drop in a server log and see which AI bots actually visit and whether your server turns any of them away.
Search
List of bots and user agents
A searchable list of bots: who runs them, what they do and how to handle them in robots.txt.
Table
AI knowledge cutoff dates
How far the knowledge of ChatGPT, Claude, Gemini and Grok reaches. Cutoff dates in one place.
Structured data won’t help if bots can’t reach your site. Check that with High Five, AI! by Maciej Chmurkowski.
Further reading
Posts on entities, Wikidata and structured data
Post
Entity, context and intent
Why keywords alone no longer cut it, and what search engines look at instead.
What next
Did the map show a mess?
One person under several inconsistent identifiers, a person mixed up with a company, profiles assigned to the wrong subject or a Wikidata record that says something different from the page. That’s the kind of thing I sort out in entity SEO work.
Want me to take a look?
Send a short brief: your site and what the map showed. I’ll tell you what is worth fixing first.
Frequently asked questions
What is an entity graph?
Is this Google’s Knowledge Graph?
Why is knowsAbout: “SEO” not an entity on the map?
Does sameAs to Wikidata connect my person to the record?
Why did the tool reject my block?
What does a “deprecated” rank mean in Wikidata?
Does Entity Observatory work on a phone?
Do you store the files or addresses I check?
Created by Ewelina Podrez-Siama, in SEO since 2009. Latest version of the tool: 0.5.1, November 3, 2026.
Wikidata data: Wikimedia Foundation, CC0 license. schema.org context: schema.org. The tool is not affiliated with either project.