Features
What the engine actually checks.
Seven capabilities, each built around a specific way people lose money on new tokens. Nothing here is a category name with a stock icon attached — every one maps to a signal the scoring engine reads.
AI rug detection
The conditions that make a rug possible, checked before you trade.
Six families of on-chain signal — liquidity, contract, holders, creator, trading and social — scored independently and weighted into one number. Conditions that make loss not a risk but a mechanism, like a live mint authority or a pool being drained mid-scan, cap the score outright rather than being averaged against a nice website.
- Mint and freeze authority state
- Sell simulation from a normal wallet
- Liquidity lock, burn and expiry
- Live removal detection
- Transfer tax and metadata mutability
Token safety scores
0-100, with every finding shown and explained.
The score is never delivered on its own. Each finding lists what was measured, what the threshold was, and why it matters in one sentence of plain language. Where a category is missing data, confidence drops and the report says so instead of quietly assuming the best.
- Per-category breakdown
- Findings ranked by severity
- Confidence based on data coverage
- Reproducible from stored inputs
Growth probability analysis
A modelled 0-100%, gated by the safety read.
Seven factors — momentum, volume, holder growth, liquidity, wallet flow, sentiment and social — each scored and weighted. A token has to be strong across several to score highly. And a structurally unsafe token cannot display an encouraging number regardless of its chart, because a good chart on a token that can be rugged is not an opportunity.
- Bullish, neutral and bearish signals listed
- Vertical 5-minute spikes discounted
- Wash-traded volume discounted
- Sparse data falls back to the base rate
Wallet intelligence
A healthy holder count that is really one wallet.
Cluster analysis groups addresses funded from a common source, so forty holders funded from the same wallet twenty minutes before launch are reported as what they are. Sniper and insider supply are tracked separately, because early automated buyers and dev-linked wallets behave differently on the way down.
- Funding-source clustering
- Sniper supply share
- Insider and dev-linked holdings
- Top holder concentration
Developer history tracking
What this wallet shipped before, and what happened to it.
Every prior launch from the creator's wallet, how many sustained value, how many ended in a removed pool, and how old the wallet actually is. A wallet with no history is reported as unknown rather than clean — absence of evidence is not a clean record, and treating it as one is how disposable launches pass.
- Prior launches and outcomes
- Previous rug count
- Wallet age and funding trail
- Linked wallet mapping
Smart money tracking
What the wallets that usually get out first are doing.
Net flow from addresses with a measured profitable track record, over one hour and twenty-four. Accumulation and distribution are reported as figures, not vibes — including when the smart money is already leaving a token that still looks strong on the chart.
- Net flow 1h and 24h
- Count of profitable wallets holding
- Buyer breadth vs transaction count
Real-time alerts
Stop watching the chart.
Rules on any watched token: safety falling below a threshold, growth probability rising above one, liquidity being removed, a whale position appearing, or a new launch matching your filter. Delivered in-app or to Telegram, with a cooldown so a token hovering on a threshold does not notify you every scan.
- Safety drop and growth spike
- Liquidity removal in progress
- Whale movement
- New launch filters
- Telegram delivery on Max
Try it on a token you already hold.
Fifteen scans a day on the free plan. No card, no trial countdown.