Free · No sign up · Nothing stored

If a machine cannot identify you, it will never recommend you.

Assistants do not browse the way you do. They lean on structured records, and when those records hold nothing for your name you get skipped in favour of somebody they can identify. This reads all four of those records at once.

Knowledge Graph Wikidata Wikipedia Google Books
Field guide

How to read what came back.

Most people run a search, see a number and have no idea whether it is good. Here is every field, what produces it, and whether it is worth caring about.

The source rowfour checks
Whether each of the four records holds anything for this name. A green Knowledge Graph light on its own is the common healthy state. Four greens is unusual and means you are a fully resolved public entity that machines can answer questions about.
Match strengthresultScore
How strongly the entity matched the exact words you typed, put on a readable scale. Searching a fuller name moves it. A high figure means Google is confident this is the primary meaning of those words. A cluster of entries all sitting at the same low figure means the opposite. It is not a trust rating and it cannot be optimised directly, whatever anyone tells you.
Entity type@type
What kind of thing Google believes this is. Person, Organization, Book, Place. If your own name returns Organization rather than Person, Google has merged you with your company, which is a real and fixable problem rather than a cosmetic one.
Entity ID/g/ or /m/
The identifier. An entry with an ID is a thing. A name with no ID is text. Identifiers starting /m/ are older records inherited from Freebase, the database Google bought in 2010, and usually indicate an entity that has existed for many years. Newer ones start /g/.
Statements and sourcedWikidata
How many facts the Wikidata record asserts and how many carry a citation. The percentage matters far more than the total. Twenty statements with two sources is a record living on borrowed time. Fourteen with nine is a record that survives scrutiny.
Knowledge Graph ID on WikidataP2671
The most telling field in the whole search and the one almost nobody checks. When Wikidata records a Google entity ID, the two systems have publicly agreed they describe the same thing. Its absence is the most common gap on otherwise decent records.
Published worksGoogle Books
Books indexed against this exact author name. An ISBN creates a permanent, structured, third-party record, which is the strongest single accelerator available to a private individual. If your books are split across two spellings of your name, this is where you will see it.
Diagnosis

Four things you might see, and what each one means.

Nearly every search lands in one of these four states, and the right next move is completely different in each.

Outcome one
One strong entry with a description
Google has resolved you. There is a real chance a panel is already showing and you have simply never checked while logged out. An identifier exists, which means there is something to claim.
Do this nextSearch your name in a private window. If a panel is there, claim it this week through the three dots on the panel itself. Claiming is free and takes minutes.
Outcome two
Several entries, all weak, none described
Fragments. Google has met your name in several disconnected contexts and built a thin partial record for each without concluding they are one person. Very common after a change of role, city or name spelling.
Do this nextStop creating new profiles. Make every existing one identical: same spelling, same photograph, same first line, all pointing at one personal site. Consolidation, not addition.
Outcome three
A strong entry, but it is not you
A namesake holds the meaning of your name. An actor, an athlete, an academic. This does not make a panel impossible, but your evidence has to be unambiguous about context: your city, your field, your organisation, beside your name everywhere.
Do this nextChoose one distinguishing form of your name and use it without exception, middle initial included. Then make sure occupation and location sit beside it on every source.
Outcome four
Nothing at all
The most common result and the least informative. This index is partial and lags the live graph, so plenty of people with healthy panels return nothing here. It may equally mean Google genuinely holds no entity for you.
Do this nextCheck in a logged-out browser first, since that settles it. If there is genuinely nothing, the work starts with one personal site and two or three independent mentions, not with a tool.
Background

What the Knowledge Graph actually is.

Google launched it in 2012 with a line that explained the whole idea: things, not strings. Before it, the search index was a very good filing system for documents. It could find pages containing the characters you typed without having any notion that those characters referred to a person who holds a job, was born somewhere and wrote a book.

Much of the original data came from Freebase, a community-built database Google acquired in 2010 and later retired, folding its contents inward. That inheritance is still visible today: identifiers beginning /m/ are Freebase-era records, while newer entities carry /g/. When Freebase closed, much of that community work migrated to Wikidata, which is a large part of why Wikidata carries the weight it does now.

The practical consequence is the part worth understanding. Ranking and being known are two different achievements. You can rank first for your own name and still not exist as an entity, which is exactly the position most senior people are in. Their site ranks. Their identity is nowhere. And because AI assistants lean on resolved entities rather than on documents, being a string rather than a thing is now the difference between being named in an answer and being absent from one.

This tool queries the public interface Google offers onto that graph. Worth knowing: Google itself now describes that interface as being replaced by a newer enterprise product, and it has always been a partial view rather than a live mirror. That is not a flaw in the tool, it is the nature of the data, which is why every result here carries the caveat openly rather than hiding it.

Why these four

What each record contributes to whether a machine can name you.

These are not four versions of the same check. Each one feeds a different part of how an assistant decides whether it knows who you are, and they carry very different weight.

Knowledge Graphgoogle entity
Whether a major search engine has resolved your name into a thing with an identifier. This is the closest available proxy for "does a machine consider this person to exist", and it is the first thing to establish.
Decisive
Wikidatastructured facts
A machine-readable statement of your occupation, employer, education and links, in a database that is openly published and widely reused. Language models draw on this kind of structure heavily, because it is unambiguous in a way that prose is not.
Decisive
Wikipediaencyclopedic prose
Strong where it exists, because it is heavily represented in training material and heavily reused elsewhere. Its absence is the norm for professionals and says almost nothing.
Strong, rare
Google Bookspublished work
A permanent third-party record tied to an author name, mirrored across catalogues worldwide. It is one of the few things an individual can create that does not decay and cannot be quietly deleted.
Durable
Your own sitenot checked here
Deliberately excluded from this check. It matters enormously for the build, and it proves nothing about whether machines have accepted your identity, because you wrote it. Including it would flatter every result.
Excluded
Social followingsnot checked here
Also excluded, and for a similar reason. Audience size is not an input to entity resolution. A million followers and no entity record is a real and common combination.
Excluded
What to do next

Four results, four different first moves.

The point of a check is the decision that follows it. These are the four patterns that come back, and what each one actually calls for.

Pattern one
All four dark
Nothing in any record. Common, and less alarming than it looks, because the public index lags. But if a logged-out Google search also shows nothing about you, this is a genuine standing start.
First moveOne page about you as a person on a domain you own, then two independent mentions. Nothing else works before this exists.
Pattern two
Knowledge Graph only
The healthiest common state. A machine has resolved you, and there is no structured record backing it up, which limits how confidently anything can describe you.
First movePublish machine-readable identity markup on your own site, then pursue a properly referenced structured record if the sourcing supports one.
Pattern three
Several weak entries
Fragments. One person, several thin records, never assembled. This is the state most likely to produce an assistant blending you with somebody else.
First moveStop creating profiles. Make every existing one identical, down to the middle initial and the photograph, and point them all at one page.
Pattern four
Strong, but not you
Somebody else owns the name. Every generic effort from here feeds their entity rather than yours, which is why people in this state work hard and see nothing move.
First moveAttach a permanent qualifier to your name everywhere. Field, city, or a middle initial used without a single exception.
Honest limits

What this check cannot tell you.

Worth stating plainly, because tools in this category routinely imply more precision than they have.

It cannot tell you what an assistant will actually say about you. There is no interface that exposes that, and any product claiming to score your standing inside a language model is inventing the number. What this checks is the underlying records those systems draw on, which is upstream of the answer rather than the answer itself.

It cannot prove absence. Google's public index is partial and lags the live graph, so an empty result means the index has not caught up, or that nothing exists. Those two situations look identical here and are completely different in practice. The only way to separate them is a logged-out search.

It cannot see non-English records. The Wikipedia check covers English only. Somebody substantial in another language will appear thinner here than they are.

And it cannot tell you whether you deserve to be named. That is decided by whether you have done things other people found worth writing down. No check, and no consultant, substitutes for that part.

Plain English

The words this field hides behind.

Entityconcept
A thing treated as an object rather than as characters. You are either an entity or you are text, and everything else follows from which one you are.
KGMID/g/11abc123
Knowledge Graph Machine ID. Google's internal reference number for an entity. Never shown on the page, retrievable through the API, and the thing you ultimately claim.
Freebase/m/ ids
The community database Google bought in 2010 and later closed. Its records seeded the Knowledge Graph, and identifiers starting /m/ still betray that origin.
WikidataQ-numbers
A structured public database of things, machine-readable, that both search engines and language models draw on. Lower bar than Wikipedia, but records still need real citations to survive.
P2671property
The Wikidata field holding a Google Knowledge Graph ID. Filled in, it means both systems have publicly agreed they describe the same thing.
resultScorematch
How strongly an entity matched the exact words typed. Relative to the query, not an absolute rating, and routinely mis-sold as a confidence or trust score.
sameAsmarkup
The list declaring which other web addresses refer to the same subject. The mechanism by which scattered profiles become one identity.
Entity homeyour site
The single canonical page unambiguously about you as a person, on a domain you own. Not a company about page and not a social profile.
Fragmentsthe failure
Several weak partial records for one person, held apart because Google has not concluded they are the same. Shows as multiple low-scoring entries with no descriptions.
Claimingverification
Proving to Google you are the subject of a panel, which earns you standing to request corrections. Requires identification and only the subject can do it.
Frequently asked

Have questions? Straight answers.

No hedging and no upsell hidden in the footnotes.

Does an empty result mean I have no Knowledge Panel?
No. This reads Google's public Knowledge Graph index, which is partial and lags the live graph badly. People with full panels, image stacks and book carousels return absolutely nothing here. A result is meaningful evidence. An empty result tells you only that this index has not caught up. Search your name in a logged-out browser to know the truth.
What is the ID that looks like /g/11abc123?
The Knowledge Graph Machine ID, Google's internal reference for a thing. If a name returns one, Google has begun treating it as an entity rather than a string of letters. It is the hook everything else hangs on, and it is what you eventually claim. You can paste one straight into the search box above.
Is match strength a Google trust rating?
No, and this gets sold dishonestly all over the internet. It comes from resultScore in Google's Knowledge Graph API, which measures how well an entity matched the exact words you typed. It moves when you change the query. Useful as a directional signal, worthless as a trust metric, and anyone presenting it as a confidence score either does not know or is hoping you do not.
Why do I see several results for one name?
Because Google is holding several unmerged fragments of one person, or genuine namesakes, and has not decided which is which. Several weak entries at identical low scores with no descriptions usually means fragments rather than different people. Merging them into one entity is real work.
Why check Wikidata, Wikipedia and Books too?
Because those are the machine-readable records that Google and the language models read when deciding what a person is. Wikidata is where an entity can be explicitly tied to a Google Knowledge Graph ID. Books create permanent third-party records against an author name. Between them they explain most of why one person has a rich panel and another has none.
Do you store what I type?
No. Google lookups pass through this site so the key stays on the server, and answers are cached briefly so repeat searches do not burn quota. Wikidata and Wikipedia are called straight from your browser. Nothing is logged against you and there is no email capture anywhere on this page.
If the result came back thin

A thin record is information, not a verdict on you.

It usually means the evidence exists but sits scattered, unlinked and unreadable by a machine. That is a fixable problem, and it is most of what the work actually consists of. If it means the evidence genuinely does not exist yet, the honest answer is to go and earn some first, and we will tell you that rather than take your money.

See how a panel gets built →