How it works
The six modules
Four modules earn trust in the number. Two are datasets that compound — they are the reason a competitor who copies the interface on day one is still months behind.
01 — Contract Scanner
What runs on real chain data today. The contract scanner below, in full. The pool reads behind liquidity, the holder reconstruction behind distribution and the bundle window, and the swap tape behind momentum — all of them on the score page, computed at the block you ask at.
What is partial. Deployer X-Ray walks the funding trail for real, but the track record it is specified as — tokens launched, graduation rate, rug rate — needs closed outcomes, so the axis caps its own confidence and says so rather than borrowing the full definition's authority.
What still runs on mock data. Smart Money and Wallet X-Ray, which are datasets that only exist once the indexer has been running. Those screens carry a mock-data marker.
The difference between a useful scanner and a checklist is simulation. Reading source tells you what the author wrote; simulating tells you what the contract does to you right now. An eth_call simulates a real buy and a real sell against current state. If the sell reverts, it is a honeypot — no interpretation required, and the hard override applies.
Alongside the simulation, each contract is checked for:
- A mint path still reachable after deploy
- Address-gating functions that can block selling
- A tax setter, and whether it is capped
- Ownership renounced or retained, and how long since deploy
- Proxy upgradeability, and whether the implementation already changed
- Max-transaction and cooldown traps
Bytecode fingerprinting
Every contract is fingerprinted against a library of known templates, both good and malicious. A contract byte-identical to one that drained something last week is flagged before anyone buys it — which is the only moment the flag is worth anything.
02 — Deployer X-Ray
A wallet's age tells you nothing on its own. What matters is where its gas came from. Every deploy is walked backwards until the funds reach an exchange withdrawal, a mixer, or a wallet already in the database.
Per deployer the database holds:
- Tokens launched, and how many graduated
- How many rugged, and the median lifespan of their tokens
- The funding trail, hop by hop, with its terminal classification
- Every other deployer resolving to the same funding root
Deployers sharing a root are one sybil cluster, not several independent actors. That is what stops a serial rugger from resetting their reputation by opening a fresh wallet: the wallet is new, the money is not.
A deployer with no history scores 50 — treated as unknown, never as safe.
03 — Bundle & Sniper Map
The first blocks after deploy are analysed to find wallets the creator funded in order to buy their own supply. The output is the share of supply held by bundles and snipers versus genuinely organic holders, drawn as clusters sized by the supply they hold and linked by their common funding source.
The cold number does more work than any amount of prose: 42% of the supply is bundled across 31 wallets from one source ends a conversation.
04 — Holder Health
Concentration matters less than its direction. A flat 30% top-10 and a 30% that was 24% yesterday are different situations, and only one of them means someone is accumulating.
- Top-10 concentration and its 24-hour drift
- Unique holder velocity, with bots filtered out
- Buy/sell pressure ratio
- Wash-trade detection, so fabricated volume cannot lift the momentum signal
05 — Smart Money Tracker
A maintained list of wallets that historically entered early on tokens that went on to graduate. Wallets are ranked by realised win rate — the share of their early entries that survived — rather than by profit, which is far easier to fake with size.
The conviction alert fires on the second tracked wallet to accumulate the same young token, not the first. One wallet is noise; two independently arriving at the same token within a short window is a signal. The alert carries the entry market cap, so the reader knows exactly how late they are.
06 — Stock-pair Native
The one module that cannot be ported to another chain. Where a token is paired with a stock token, the vault is checked for whether it is actually accumulating rather than merely announced, at what rate, and how fast supply is being burned.
An announced vault with a zero balance is a different asset from one buying daily, and a generic multichain screener has no column for the distinction.