SOL balance
Native SOL held by the wallet at scoring time.
Medusa whitepaper
How the wallet credit score is calculated.
Medusa converts public Solana Devnet wallet behavior into a readable borrower reputation score. The current model is `trust-rpc-0.2` and produces a score from 8 to 98 points.
Data inputs
Token accounts
SPL token accounts, with stronger weight for non-zero balances.
Transaction sample
Recent signatures fetched from Solana RPC and checked for success or failure.
Wallet age
Observed days between the oldest sampled activity and the current score request.
Active days
Unique days with activity inside the sampled transaction window.
Protocol touches
Known DeFi or protocol interactions found in the sampled wallet activity.
Repayment behavior
35 pts
round(successRate x 35)
A higher successful transaction ratio increases the strongest score component.
Liquidity resilience
30 pts
round(min(SOL, 10) x 1.5) + min(nonZeroTokenAccounts x 3, 15)
Rewards available SOL and diversified non-zero token balances, capped to prevent oversized wallets from dominating.
Wallet age
25 pts
round(min(observedAgeDays, 365) / 14.6)
Older observed history improves confidence, with one year treated as the useful cap.
Protocol activity
18 pts
round(min(sampledTransactions, 50) / 3.2) + min(activeDays, 6)
Rewards steady, repeated wallet use across multiple days.
Protocol bonus
6 pts
round(protocolInteractions / 4)
Adds a small bonus for known protocol participation without overpowering core wallet behavior.
Risk flags
12 pts penalty pool
round((1 - successRate) x 8) + failedTxPenalty + lowSamplePenalty
Failed transactions and thin history create a penalty that is partially subtracted from the score.
Final score equation
score = clamp(repayment + liquidity + walletAge + activity + protocolBonus - floor(risk / 2), 8, 98)
The model intentionally caps the score at 98 because this version is based on public RPC observations, not a complete off-chain identity, underwriting, or repayment registry.
Score bands
Prime
82-98
Strong history, stronger liquidity, and low observed risk.
Reliable
68-81
Enough wallet evidence for risk-adjusted DeFi credit review.
Emerging
52-67
Some useful activity, but history or liquidity is still developing.
High Risk
8-51
Thin, failed, or uneven activity. Conservative review recommended.
Badge mapping
LV4
Prime
Minimum score: 82+
LV3
Reliable
Minimum score: 68+
LV2
Emerging
Minimum score: 52+
LV1
High Risk
Minimum score: 8+
Confidence and limitations
API confidence is calculated as `clamp(45 + sampledTransactions + activeDays x 2, 45, 92)`. The score can be limited by RPC availability, small transaction samples, test-wallet behavior, spam, failed transactions, or wallets with meaningful activity outside the sampled window. Medusa should be used as a transparent reputation signal, not as automatic loan approval.