SOAR SCAN · dev graph
Our graph links deployer wallets by following money, shared identifiers and shared behaviour. The honest way to show whether that works is not to display what it found on its own — it is to hand it addresses somebody else published first, and report what it does with them. Including the ones it cannot connect, which are the more useful half.
The rule this page runs on. We never say an address belongs to a scammer. That claim is not something the chain can tell us — it comes from trusting an article, and if the article is wrong we would have published a permanent accusation against a real person. So the claim stays the researcher's, cited with a link you can check. What we add is separate and is ours: whether the chain confirms the behaviour they described, and whether our graph reaches the same connections without being told about them.
Three, and that is not many. They are labelled by kind above, because a peer-reviewed dataset and a blog post are different evidence and flattening them into one word would be its own dishonesty.
One source naming an address is a report, not a settled fact, and the terminal says exactly that when it shows you one. Two independent sources naming the same address is a different weight, and it is marked differently. Right now no address in here has two — so nothing in here has reached that bar, and we say so rather than letting the page imply otherwise.
One more distinction that matters: the academic paper names tokens, not wallets. The deployer wallets under it were resolved by us from the chain, and they are labelled derived — reporting them as the authors' finding would put words in their mouth.
A case our graph does not connect is not a bug we are hiding. There are three reasons a trail ends, and all three are printed above when they happen: the wallet has more history than we can read, the money came from an exchange hot wallet shared with a million strangers, or the connection is real but sits further away than the three hops we follow.
The clearest example we know of is one we would fail. Bubblemaps connected the WOLF memecoin to LIBRA by following funding across seventeen addresses and five cross-chain transfers. Our method stops at three hops on Solana alone, so it would not have found that, and no amount of confidence on this page changes it. (The Block, on Bubblemaps' analysis)
Because a list is an accusation, and ours would eventually be wrong about somebody. The database physically refuses to store a case without a source URL — not as a policy, as a constraint in the code. What our own graph reports instead is what the chain shows: this wallet deployed four tokens, none of them still trade. Nobody can be defamed by a measurement.