The methodology behind every grade.
When we say “yield research you can audit,” this page is the thing you audit. The signals, the weights, and the thresholds below are the model's real constants - the same code that grades every pool renders this page.
What a grade is made of
Every pool earns a 0–100 score - a weighted blend of four transparent signals. The grade also names its weakest signal, so an A-grade pool that is strong everywhere except reward quality tells you exactly where to look first.
How much value backs the pool, and how deep it is.
How steady the yield has been, not just its headline.
How much of the APY is real yield vs. incentive emissions.
How much of the picture the market feed actually shows.
Thresholds are fixed, not curved - a pool is never graded on its neighbors. When the model changes, that change is published like every snapshot, so the methodology's own history is auditable too.
Where the data comes from - and where it stops
Pool data is sourced from public DeFiLlama feeds and refreshed on a live scan. Each scan is snapshotted, so Litmus accumulates its own APY and TVL history over time; pool detail pages also pull per-pool history directly from DeFiLlama for 30- and 90-day context.
Just as important is what a grade is not: it does not review smart-contract code, audit a protocol's team, or see anything the public market feed can't. A grade is not an audit. Data can be delayed or incomplete - always confirm on the protocol itself.
Every grade, on the record, before the outcome
Every Monday, an automated job grades every live pool and publishes the full list as a dated, timestamped snapshot. Each snapshot is append-only and written beforeanyone knows how those pools perform - the opposite of a cherry-picked backtest. To make that tamper-proof rather than just “trust us,” each snapshot's hash is anchored on the Base blockchain - an immutable public ledger (launching soon). Every snapshot records, for every pool:
poolIdstable DeFiLlama pool identifierproject / chain / symbolwhich protocol, network, and assetapythe APY at grading timetvlUsdpool TVL at grading timegrade / scorethe Safety Grade letter and its 0–100 scoreSee the running history, the grade-mix charts, and the on-chain verification status on the track record page.
Alerts in plain language
Alerts are built in plain language - pick what should trigger one and Litmus watches for it on every scan: APY rising above or dropping below your threshold, a sudden APY fall, TVL draining, or rewards making up too much of the yield. Alerts fire only on a real match - never spam - by email or Telegram. Separately, the weekly digest emails you every Sunday with your watchlist's APY and grade moves; one toggle turns it off.
Straight answers
Is a Safety Grade an audit?
No. It is a published-methodology opinion computed from public market data. It does not review contract code or teams, and it is not a certification of safety.
Does Litmus hold my funds?
No. It is non-custodial and never connects a wallet or executes transactions. It is research and alerting only.
Is this financial advice?
No. Litmus provides research and signals; every decision and its consequences are yours.
Where does the data come from?
Public DeFiLlama feeds, plus Litmus’s own accumulating snapshot history.
Can past grades be edited?
No. Each weekly snapshot is recorded before its outcome is known and is append-only; on-chain anchoring on Base (launching soon) makes any change independently detectable. That is the point of the record.
Disclaimers
Onchain yield carries significant risk, including total loss of capital. Safety Grades and scores are estimates, not assurances. Verify every opportunity with the underlying protocol before moving funds. See our Terms, Privacy, and Disclaimer for the full picture.