
SEMI-LIQUID
Real depth, boring price action, an oracle nobody argues about. It still gets a sealed barrel, because tiering is not trust.
- Max LTV
- 55–65%
- Hazard premium
- +0 to 3%
Blue-chip lending markets list a dozen assets and refuse the rest. We take the rest. Every radioactive asset gets sealed in its own barrel — its own oracle, its own caps, its own liquidations — so a loss in one cannot reach another.
Isolation limits contagion. It does not eliminate your loss on a given barrel. Figures on this page are testnet simulations, not returns.
Aave lists a few dozen assets. Everything else — the long tail, the majority of what actually exists on-chain — sits idle, because one bad listing can create bad debt that spreads to the whole pool. So nobody lists it. Rationally.
Our bet is narrow and testable: you can safely accept toxic collateral if you isolate it completely.
This is the whole product, and it is a mathematical claim, not a marketing one: if any code path can socialise a barrel's loss beyond that barrel, we treat it as a bug.
Each market is a single collateral / borrow pair with its own oracle, LTV, liquidation threshold, caps and rate model. Nothing is shared. There is no pooled reserve to drain.
If liquidation proceeds fall short, the shortfall is absorbed by that barrel's backstop and its lenders. It never reaches another barrel, and never reaches the treasury.
Tiering changes parameters, never isolation. The tamest collateral we hold is sealed exactly as hard as the worst one.
Six assets have been through the listing process. Four are live, one waits on an oracle we can trust, and one is sealed permanently. That ratio is the point.
The full inventory — including the queued and the permanently sealed — lives in the app, with positions, health factors and the backstop.
Launch the app☢ Testnet simulation. Rates and balances shown are illustrative and are not an offer, a forecast, or a promise of return. Supplying to any barrel risks the full loss of your deposit.
A barrel produces revenue from three places. Once lenders are made whole, the surplus — the Overflow — goes to the people who carried the risk: the backstoppers staked against that specific barrel, and $FLLT stakers.
The hazard premium. Toxic collateral borrows at a rate that pays lenders for holding real risk.
Liquidation bonuses paid by borrowers who let a position drift, collected inside the barrel.
A cut of the above, swept per barrel. Nothing is minted. Nothing is subsidised.
of distributions are paid in the borrow asset. We will not print $FLLT and call it yield. If a barrel earns nothing this week, it distributes nothing this week — and the dashboard will say so.

Real depth, boring price action, an oracle nobody argues about. It still gets a sealed barrel, because tiering is not trust.

Thin books, weird price surfaces, LP positions priced from their legs. Tight caps, fast liquidations, a backstop that gets paid to be there.

The stuff every other market refuses outright. Micro caps. If it goes, it goes fast — and it goes alone, inside its own barrel.
Barrels are curated, not permissionless — for now. Three operators sign off on every listing, every parameter and every oracle. Rejections are published alongside listings.

“My job is mostly saying no. I have rejected four assets for every one that got a barrel, and the rejections are the part I am proud of.”

“A barrel is exactly as safe as its price feed. If I cannot break the feed in a simulation, it ships. If I can, the asset waits — however good the yield looks.”

“We model who actually shows up to buy the collateral at 3am on a red day. If the answer is nobody, the cap goes down until the answer changes.”
Isolation caps how far a loss travels. It does not cap your loss. If a barrel takes bad debt beyond its backstop, its lenders absorb the rest.
Oracle manipulation is the number one attack on any money market, and it is worst exactly where we operate. We use robust feeds, TWAPs, staleness and deviation breakers — and we refuse assets we cannot price safely.
If nobody bids at liquidation, the barrel eats the difference. That is why caps are small, why liquidations fire early, and why the backstop exists.
Nothing here is live on mainnet. Nothing here is an offer, an invitation, or financial advice. Smart contracts fail; assume this one can too.