DEBB analyzes any Solana token in real-time: deployer history, holder concentration, contract permissions, and launch behavior. One REST call, built to embed in your bot, launchpad, or wallet — not just check tokens by hand.
Free: Telegram bot + 10 API calls/day.
DEBB runs five on-chain checks in parallel and returns a risk score in seconds. No setup, no SDK. Just a REST call.
POST a Solana token address to /v1/score with your API key. Any SPL token, any time.
DEBB runs 5 on-chain checks in parallel: deployer history, holder data, contract flags, launch pattern, LP status.
An XGBoost model trained on labeled rug outcomes returns a 0–100 score, a verdict, and a signal quality rating. Cached results return right away; fresh scans usually finish within 12s.
The score comes from an ML model with six on-chain inputs, trained on labeled rug outcomes. Holder share is shown for context and is not scored. The model estimates risk; it can't guarantee a token is safe.
Tracks how many previous tokens from the same wallet ended in a rug. A deployer's past rugs raise the risk score.
Counts the danger-level risks RugCheck reports for the token (up to 5). More flags means more ways the setup can work against holders.
Measures how broadly the token is held. A single wallet dominating supply means a single sell can collapse the price. We flag when the top 10 holders own most of the supply.
Detects coordinated insider buying in the same block as launch. Organic launches look different from insider-first accumulation, and this signal looks for that difference.
Whether 95%+ of the pool's Raydium LP tokens are burned (BURNED). UNLOCKED means the deployer can pull liquidity at any time. N/A means no Raydium pool was found or the check failed — expect this on most pump.fun tokens.
DEBB scores tokens with an ML model trained on labeled rug outcomes, and doesn't bluff when the data is thin.
Every score includes a signal_quality field: high, moderate, or low, based on how many of the 5 checks returned real data, plus the count itself. When data is thin, we say so.
Because one missed rug costs more than a cautious score.
A single REST endpoint. JSON in, risk score out. No SDK required. Integrate into any trading bot, launchpad, or analytics platform. DEBB scores every token with a trained ML model, and every score ships with a signal_quality field so you can see exactly how much data backed the call.
curl -X POST https://api.debb.fyi/v1/score \
-H "x-api-key: debb_your_key_here" \
-H "Content-Type: application/json" \
-d '{"token": "7xKXtg2CW87d97TXJSDpbD5jBkheTqA83TZRuJosgAsU", "chain": "solana"}'{
"score": 14,
"verdict": "danger",
"signal_quality": "high",
"signals_measured": 4,
"signals": {
"deployer_rug_rate": 0.78,
"bundled_launch": true,
"contract_flags": 2,
"holder_top10_pct": 0.21,
"lp_locked": false
},
"latency_ms": 847
}x-api-key header✓ Recent results are cached (minutes for new tokens, up to 24h for older ones)✓ Monthly subscription, cancel anytime✓ Fresh scans usually finish within 12sMonthly plans, cancel anytime. Or start free via Telegram and upgrade when volume demands it.
Cancel anytime
Cancel anytime
Cancel anytime
Not ready to commit? Free: Telegram bot + 10 API calls/day.
Embed real-time, ML-scored deployer risk intelligence into your bot, launchpad, or wallet with one API call — or check a token yourself first via Telegram.
The Rug Report
Built for traders who've been burned before.
Decide on verdict, not score. The score each verdict needs differs by the creator's history: a brand-new creator needs a higher score to rate lower risk than a known one. The headline field states the bar, e.g. "RISKY · 84/100 · new creator: Lower risk needs 88+". A check that returned no data comes back as null and is listed in unknown_signals. When at most 1 of the 5 checks returned data and nothing raised a warning, insufficient_data is true: treat the token as unrated (its verdict and score stay for older clients).