onchain_smart_wallets
onchain_smart_wallets
Top-scored on-chain wallets surfaced by the FalsifyLab live Solana copy-trading bot. Composite score derived from five weighted sub-scores (profitability, consistency, risk, copyability, behavior_quality) calibrated against actual copy-trade PnL correlations.
Consistency and behavior_quality are weighted highest (rho +0.76 and +0.755 to copy-trade PnL respectively). Profitability and risk dims are de-weighted because high-PnL-on-paper wallets bleed under copy-trade latency tax.
Signature
onchain_smart_wallets({ min_score?: number, // composite floor 0.0-1.0, default 0.5 status?: "MONITORED" | "CANDIDATE" | "ACTIVE" | "DEMOTED", chain?: "solana" | "ethereum", limit?: number, // 1-50, default 10})Returns
{ "count": 3, "items": [ { "address": "4DmWwzuorfwXSiCFamtYmmZ6AXMDPZCsu3J6o26RvE1N", "chain": "solana", "status": "MONITORED", "score_total": 0.8847, "score_breakdown": { "profitability": 0.6226, "consistency": 0.9255, "risk": 0.8, "copyability": 0.6428, "behavior_quality": 0.9503 }, "provider": "helius", "scan_window_days": 30, "confidence_level": 0.94, "scored_at": "2026-05-14T01:54:17Z", "explorer_url": "https://solscan.io/account/4DmWwzuorf…" } ], "scoring_weights": { "profitability": 0.10, "consistency": 0.40, "risk": 0.05, "copyability": 0.05, "behavior_quality": 0.40 }}Status values
MONITORED— actively scored, eligible for copy-trade signalCANDIDATE— under evaluation, insufficient history for full classificationACTIVE— same as MONITORED but actively traded againstDEMOTED— failed recent performance threshold; retained for re-evaluation
Example agent prompt
“Run onchain_smart_wallets with min_score=0.7 and status=MONITORED. List the wallets sorted by consistency desc. For each, explain what its behavior_quality score tells me.”
Why this exists
Most wallet-tracking tools rank by PnL alone. That misses two failure modes:
- Latency tax: top-PnL wallets often execute in pre-confirmed mempool tricks that retail copy-traders can’t replicate; their on-paper PnL doesn’t survive a copy.
- Pattern volatility: wallets that pump 10x then bleed don’t make consistent copy-trades — they roulette.
The composite score is calibrated from a -0.63 → +0.76 correlation reversal audit on the FalsifyLab copy-trade bot. Weights pin the most predictive sub-scores (consistency + behavior_quality) at 40% each.
Free tier
24h cached, 10 results minimum_score 0.5 floor. Covers “show me the top scored wallets right now” queries.
Pro tier
5-minute refresh, 50 results, full sub-score history (90d), filter by chain or status or score floor.
Caveats
- This is research data, not financial advice. Wallet behavior on-chain is public, but past performance does not predict future copy-trade profitability — copy-trade latency, fee structure, and signal decay all matter.
- Status transitions can lag 24h. A wallet flagged DEMOTED may have been re-eligible for hours.
- Chain coverage is currently Solana-heavy. Ethereum wallets are tracked but with smaller sample sizes.
See also
- hl_vault_leaderboard — Hyperliquid vault ranking with similar composite scoring
- polymarket_whale_positions — prediction-market wallet tracking
- confluence_today — when this wallet’s activity stacks with other signals