Live, signed, hourly
The Honesty Index
Every synthetic dollar publishes an APY. Almost none publish what their holders actually earned. Those are different numbers, and the distance between them is the only interesting thing in this sector.
So we read it. For everyone, hourly, from the one figure a synthetic dollar cannot spin: the share price of its own vault, which is public, which settles its own redemptions, and which anyone can re-read. We put ourselves on the board first.
What it says right now
Of the 3 competitors we can compare on a like for like basis this hour, 3 land within half a percentage point of their own published number. 2 more publish no yield figure at all, because they are in recovery or wound down, and say so. On the one axis this page measures, the field is largely honest.
The largest gap between advertised and realized on this board is Kerne's, at about 13 points. That is us. We built the scoreboard and we are the worst row on it. The rest of this page explains why, without excusing it.
| Protocol | Advertised, in their own words | Realized, from the chain | Gap | Size, in its own asset | Status |
|---|---|---|---|---|---|
Kerneus skUSD base | 12.91% skUSD APY (modeled) (modeled) read 2026-07-15 from their own source | 0.33% share price 1 to 1.000098667771066974 over 10.9134 days, since this vault started accruing | +12.58 pp | 1.0k kUSD | Live |
Cap stcUSD ethereum | 5.05% 7 day average USD APY (7 day average) read 2026-07-14 from their own source | 5.24% share price 1.066204644622994646 to 1.070692086709928203 over 30 days | Not comparable | 52.12m cUSD | Live |
Ethena sUSDe ethereum | 3.72% 30 day average sUSDe yield (30 day average) read 2026-07-14 from their own source | 3.75% share price 1.235516427822250771 to 1.239263208439322878 over 30 days | -0.03 pp | 1.57b USDe | Live |
Falcon Finance sUSDf ethereum | 4.86% sUSDf APY (7 day average) read 2026-07-14 from their own source | 5.12% share price 1.120663507044044095 to 1.125274931710752759 over 30 days | Not comparable | 69.40m USDf | Live |
Neutrl sNUSD ethereum | 4.30% 30 day average APY (30 day average) read 2026-07-14 from their own source | 4.34% share price 1.056621410263071991 to 1.0603191540042666 over 30 days | -0.04 pp | 39.48m NUSD | Live |
Sky sUSDS ethereum | 3.60% Sky Savings Rate (current rate) read 2026-07-14 from their own source | 3.60% share price 1.099824807078569002 to 1.103026526466449201 over 30 days | 0.00 pp | 4.77b USDS | Live |
Resolv wstUSR ethereum | None published. This protocol currently advertises no yield figure at all. | 0.00% share price 1.132966158918315225 to 1.132966158918315225 over 30 days | No claim | 2.09m USR | In recovery |
Elixir sdeUSD ethereum | None published. This protocol currently advertises no yield figure at all. | 0.00% share price 1.07056672845499187 to 1.07056672845499187 over 30 days | No claim | 21.81m deUSD | Sunset |
A positive gap means the protocol advertised more than its holders realized. A negative gap means it realized more than it advertised. Where a protocol publishes a 7 day average and our realized column is a 30 day window, we state no gap rather than manufacture one out of a mismatch of periods. The size column is each vault's own balance, in its own asset, and it is not scored: it is here so that a reader can see how small Kerne is next to the protocols it is scoring, which is a fair thing to hold against us.
Every row, in context.
Kerne skUSD
Live, and the worst gap on this page. skUSD v2 was deployed 2026-07-03 with its share price reset to exactly 1.000000, so its window is shorter than the 30 days used for every other row and is labelled as such. Read the realized figure with care, including against us: it derives from a single 0.10 kUSD distribution on 2026-07-08, which the deployment registry records as a plumbing smoke test proving the strategist topology works, not strategy carry. Annualizing one test transfer is annualizing a test transfer. Absent further distributions the figure decays daily toward zero. Kerne has not yet paid meaningful realized yield, and our published APY is a model, not a record.
Cap stcUSD
Live and accruing. Cap publishes a 7 day average and it is close to what we compute over 30 days. Note for anyone cross checking this row against an aggregator: DefiLlama lists a Cap pool near 7 percent on a different basis, which is not the figure Cap advertises. We score the protocol against its own number.
This protocol publishes a 7 day average and our realized column is a 30 day window, so the two are not directly comparable. We show both and state no gap rather than manufacture one out of a unit mismatch.
Ethena sUSDe
Live and accruing. Share price last moved minutes before this reading. Ethena publishes a current yield and a 30 day average that both track what we compute on chain. It also publishes a since inception average near 11 percent, roughly three times the current rate; that figure is correctly labelled on their surface and is shown here with its label intact.
Other figures this protocol publishes
- 3.75% sUSDe staking yield (current rate)
- 10.97% sUSDe yield since inception (since inception)
Falcon Finance sUSDf
Live and accruing, with weekly third party reserve attestations and a protocol backing ratio above par. One inconsistency on their own surfaces is worth naming precisely: the classic Earn page still carries the title "Earn Double-Digit APY on USDf Synthetic Dollars" and a description promising base yield above 10 percent, while Falcon itself renders 4.86 percent today. That is stale marketing copy contradicted by their own live figure, not a claim about current yield.
This protocol publishes a 7 day average and our realized column is a 30 day window, so the two are not directly comparable. We show both and state no gap rather than manufacture one out of a unit mismatch.
Neutrl sNUSD
Live and accruing. Neutrl labels the basis on every APY it publishes, which is better discipline than most of this sector manages, and its own 30 day average lands within a few basis points of what we compute independently on chain. It leads its homepage with a since inception figure of 8.9 percent, roughly twice its current 7 day rate. That label is present and accurate on their own surface, so the figure is shown here as context and not as a claim about what a holder earns today.
Other figures this protocol publishes
- 4.10% 7 day average APY (7 day average)
- 8.90% Since inception APY (since inception)
Sky sUSDS
Live and accruing. The Sky Savings Rate is a governance set rate rather than a strategy return, which is why advertised and realized agree so exactly. That is a structural property of the design, not a virtue Sky has and others lack, but it does mean a holder gets what the label says.
Resolv wstUSR
In recovery, and publishing no yield figure at all. Resolv disclosed a security incident involving unauthorized minting of USR and states that most protocol operations are paused. Its own staking APR endpoints now return "Deprecated: staking APR is not available", so there is no advertised number to compare against. A recovery framework is running with claims open until 2026-08-26. The share price has not moved since 2026-03-22, which is the date of the incident named in Resolv own postmortem. Reading this row as deception would be wrong: a paused protocol that stops accruing and stops advertising is behaving correctly.
Read the realized column for this row with care. A share price is denominated in the underlying, and the underlying here is USR, which trades near 0.14 dollars rather than 1. wstUSR holding steady in USR terms says nothing about the dollar outcome for a holder.
Elixir sdeUSD
Sunset, and publishing no yield figure. Elixir own documentation states that deUSD "holds no value and the asset has been sunset", and directs holders to 1 to 1 USDC claims. Its marketing site now renders a single notice stating that the Elixir Network has shut down. The share price has not moved since 2025-11-05, which sits one day after Stream Finance disclosed a loss of roughly 93 million dollars; Elixir held a majority of deUSD collateral with Stream and announced the sunset days later. The frozen share price here is a record of that, not a hidden fee.
How each number is produced.
The realized column is not an estimate, a model, or a survey. For each vault we read its own accounting, twice, and divide.
- Ask the vault for its decimals() and its asset(). We never assume 18. Kerne's own skUSD reports 24, and a reader who assumed 18 would compute a share price of 0.000001 and publish nonsense about us.
- Find the block whose timestamp is closest to 30 days before the current block, by binary searching real block timestamps. We do not assume a block interval; an assumed interval silently shifts the window by hours and every annualized number with it.
- Read convertToAssets(10**decimals) at both blocks and divide each by 10**asset_decimals. That is the share price, in units of the underlying, exactly as the vault itself would settle a redemption.
- Annualize the ratio over the actual elapsed seconds between the two blocks, not the nominal 30 days.
Where this measurement stops being fair, stated plainly.
- A share price is denominated in the underlying, not in dollars. A vault can grow steadily in units of an asset that is itself broken. That is why Resolv's row carries a peg note: wstUSR is measured in USR, and USR does not trade at a dollar.
- Windows differ. Our realized column is 30 days. Several protocols advertise a 7 day average. Those are different periods and we refuse to subtract one from the other.
- Annualizing a short window exaggerates, and here it exaggerates in our own favour. Kerne's vault is younger than the window, so its realized figure scales a single test distribution up from a handful of days rather than a full month. A longer window would spread that one transfer across more time and pull the number down, not up. We flag the clamped window on our own row for exactly this reason.
- A since inception average is not current yield. Ethena and Neutrl both publish one, both label it correctly, and we show it as context only. It is never the headline and never the basis of a gap.
- This page does not score the hedge. The hedge leg is user-unverifiable for every delta-neutral protocol here, including Kerne, whose Hyperliquid leg is self-reported and signature-bound only. We score yield, which is verifiable on chain. We do not score what we cannot see, and neither should anyone reading this.
- The advertised column is manual and dated. The realized column is live and hourly. Scraping eight marketing surfaces on a timer produces confident wrong numbers, because most of them are JavaScript shells that serve a placeholder zero to a naive fetch. So a human reads each figure, quotes it verbatim, and dates it. If a protocol changes its number and our row is stale, that is our bug, and the date is there so you can catch us.
Check us. The whole point is that you can.
Every input is public chain data. Nothing here requires our permission, our API key, or our good faith. Take any row and recompute it:
# Ethena sUSDe, current share price cast call 0x9D39A5DE30e57443BfF2A8307A4256c8797A3497 \ "convertToAssets(uint256)(uint256)" 1000000000000000000 # the same vault, 30 days earlier (any archive RPC) cast call 0x9D39A5DE30e57443BfF2A8307A4256c8797A3497 \ "convertToAssets(uint256)(uint256)" 1000000000000000000 \ --block <from_block> --rpc-url https://eth.drpc.org # our full table, including every block number we used curl https://kerne.fi/api/honesty-index
The JSON carries the from and to block for every row, so you are not re-deriving our window, you are checking our arithmetic inside it. If you get a different answer, we want to know: tell us and we will correct the row.
Each snapshot is signed by 0x84949170e0ad0f9bd8686a4aa4922c10f5fdc4ea. The signature binds us, not the protocols listed: it is what stops us from quietly revising what we said about a competitor, or about ourselves, after the fact. Verify it the same way you verify our reserves, at /verify.
What Kerne still owes.
- Realized yield. Our published APY is a model at target leverage. Our realized number comes from one 0.10 kUSD distribution on 2026-07-08 that the deployment registry itself records as a plumbing test. Until Kerne distributes real strategy yield, our row is the weakest on this page and should be read that way.
- Scale. Kerne is small. The protocols above us are between four and six orders of magnitude larger. Being honest about a tiny number is not the same achievement as running a large one well.
- An audit. Fieldwork by Hexens is underway. Until the report is public, we are a pre-audit protocol and the other rows here are not.
- The hedge. Ours is as unverifiable to you as everyone else's. We publish a signed attestation of it, which is better than nothing and is not proof.
- A longer window. skUSD is younger than 30 days, so its row is measured over a shorter period than everyone else's. That flatters nobody, but it does mean the comparison is not yet like for like, and we say so on the row.
This is one axis. Here are the others.
This page scores yield honesty and nothing else. A synthetic dollar can pay exactly what it advertises and still be opaque about what backs it, and the reverse is also true. The other axes live on their own pages, deliberately.
- Reserve transparency scorecard : can a holder recompute the reserves themselves, with no attestor in the trust path?
- Attestation versus recomputation : the trust models, mapped on a spectrum.
- The failure ledger : what actually broke, and what was legible before it broke.
- Kerne proof of reserves : our own reserves, signed and recomputable.