A credit score for any EVM wallet.
One number from 300 to 1000, a range band your policy can key on, and six factor ratings that say why. Built from repayments and liquidations on Aave, Compound and Morpho, plus how the wallet behaves across ten chains.
{
"model_version": "andromeda_1.4",
"score": 862,
"decile": 7,
"range": "very_good",
"factors": [
{ "label": "Borrowing History", "rating": "excellent" },
{ "label": "Wallet Composition", "rating": "good" },
{ "label": "Wallet Health", "rating": "very_good" },
{ "label": "Interactions", "rating": "very_good" },
{ "label": "Trust", "rating": "good" },
{ "label": "New Credit", "rating": "good" }
]
}Five bands, fixed thresholds.
Every response carries a range and a decile alongside the number, so a credit policy can key on a band without re-deriving it.
- Low300–639+
- Fair640–749+
- Good750–839+
- Very Good840–919+
- Excellent920–1000+
Six reasons behind every number.
Each factor comes back rated from Low to Excellent. You see why a wallet scored the way it did; the weights stay private, so a borrower cannot work the score.
- Borrowing History
- How the wallet has borrowed and repaid: repayments, liquidations and their record over time.
- sample: Excellent
- Wallet Composition
- The mix of what the wallet holds, from stablecoins to volatile and illiquid assets.
- sample: Good
- Wallet Health
- Overall financial standing and the absence of adverse events.
- sample: Very Good
- Interactions
- How broadly and steadily the wallet engages with onchain protocols.
- sample: Very Good
- Trust
- Identity attestations and how long the wallet has been active.
- sample: Good
- New Credit
- Recent originations and the pace at which new debt is taken on.
- sample: Good
Read from lending events, not guessed from balances.
Borrowing history comes from the events themselves: every borrow, repayment and liquidation on Aave V2 and V3, Compound V2 and V3, and Morpho Blue. Wallets that have never borrowed are still scored, from how they behave across all ten chains.
Every model release must rank a fixed set of borrower archetypes the way an analyst would before it ships. On the demo call we score wallets whose outcomes you already know.
Test it on wallets you know- Ethereumborrowing + behaviour
- Baseborrowing + behaviour
- Arbitrumbehaviour
- Optimismbehaviour
- Polygonbehaviour
- Avalanchebehaviour
- BNB Chainbehaviour
- Celobehaviour
- Lineabehaviour
- Robinhood Chainbehaviour
See your own wallets scored.
A 30-minute call with live scoring on wallets you choose, and the factor reasons behind every number. No commitment.
Building an AI agent? Get an API key