No Self-Reported Data
Every metric derives from on-chain state or audited protocol APIs. We verify share price appreciation directly from observed state, not from projected yields or protocol marketing numbers.
Stacked TVL by protocol (top book) · click legend to isolate
Active loans and TVL-weighted utilization · gaps are days the platform could not measure
Vault tier upgrades and downgrades · counts, not TVL
How scored vault capital concentrates across integrated protocol modules.
How scored vault capital is distributed across covered EVM networks.
Largest curators by scored vault TVL.
Click a row for detail.
TVL and vault mix across Prime, Core, and Edge. Click a slice to open that tier.
Latest risk and security signals.
Severity: Warning · 1h ago
Severity: Warning · 1h ago
Severity: Warning · 2h ago
Coverage counts. Combined AUM withheld pending product registry.
Each protocol is integrated through a dedicated adapter built to extract share price accrual, reward decomposition, and on-chain event streams in a format consistent with our risk framework. Coverage spans Ethereum, Base, Plasma, Arbitrum, Monad, Avalanche, Polygon, Optimism, Stable, and Plume.
V3 supply markets (incl. Lido and Horizon) on Ethereum, Base, Arbitrum, Optimism, Polygon, Avalanche, Monad, and Plasma.
Hub-and-spoke liquidity markets (Core, Prime, Plus) on Ethereum and Avalanche.
Curated lending vaults (v1, v2) with isolated risk markets on Ethereum, Base, Arbitrum, Monad, and Stable.
SparkLend (v1) and Spark Savings (v2: sUSDS, sUSDC, spUSDC, spUSDT) on Ethereum, Base, and Arbitrum.
Compound V3 (Comet) supply-side lending on Ethereum, Base, and Arbitrum.
AMM liquidity provision across V2, V3, and V4 on Ethereum, Base, and Arbitrum.
V2 and V3 vaults on Ethereum, Base, Arbitrum, Optimism, and Polygon.

V2 auto-compounding vaults on Ethereum, Base, Arbitrum, Optimism, Polygon, and Avalanche.
Nest V1 curated yield vaults on Plume and Ethereum.
Your protocol here?
Contact us →Scored, monitored, and accountable from observed on-chain state, not self-reported marketing inputs.
Every metric derives from on-chain state or audited protocol APIs. We verify share price appreciation directly from observed state, not from projected yields or protocol marketing numbers.
Each vault is scored across four independent risk vectors: asset composition, platform and strategy, control and governance, and history. The framework is deterministic: the same inputs always produce the same score.
APY is calculated from historical share price changes, reflecting actual depositor returns net of fees and slippage. Forward-looking estimates are labeled as such when shown.
Scoring criteria, vector weights, and tier thresholds are publicly documented. We invite scrutiny. If the model is wrong, we want to know.
Vaults are classified into three tiers from a weighted composite score across asset quality, platform and strategy, control and governance, and history. The tiers describe a relative risk spectrum. They are not an investment recommendation.
Strong observed controls and evidence quality relative to tracked vaults: mature codebases, meaningful audit coverage, and tighter governance constraints. This is the lowest-risk segment of on-chain yield in our framework.
A moderate risk profile with meaningful constraints or uncertainty. Typical cases include audited protocols with shorter track records, more permissive governance, or mixed vector scores.
An elevated or unresolved risk profile. Common drivers include newer code, limited audit coverage, concentrated admin powers, or weak recent history. Higher yield often accompanies higher tail risk.
Illustrative breakdown: same four vectors as vault detail pages
Illustrative scoring for a Prime-tier lending vault
Scores are derived from measurable on-chain state. Team reputation, social following, and marketing spend are not inputs.
The same methodology applies to every protocol. Listing fees fund coverage and distribution; they do not change scores, vector weights, or tier placement.
Risk scores update automatically as on-chain conditions change. Incident detection triggers immediate score adjustments.
Every vector weight, scoring threshold, and tier boundary is publicly documented. Challenge the model, not the output.
Missing data receives the worst-case assumption. An unaudited protocol scores zero on audit density rather than a neutral pending-review placeholder.
Constraints we impose on scoring and coverage so commercial relationships cannot rewrite risk output.
Filter by risk tier, chain, protocol, and asset type, then compare risk-adjusted returns across 800+ scored vaults.