Forge

Pool-based lending on QOM with a single QOM lender pool serving two loan types: a standard Direct Loan (deposit ZAP, borrow QOM, kinked-IRM, partial liquidation) and a Bootstrap Loan that turns a QOM borrow into a self-deleveraging staking position via QNS Staking lien hooks. Lenders see one APY across both loan types: $R_l(U) = R_b(U)\, U\, (1-\phi)$ with a proposed $\phi = 25\%$ reserve factor. Direct is the conventional Aave/Compound-shape primitive; Bootstrap is the novel one — a debt that retires itself in finite time when the cash-flow-to-debt ratio $\gamma > r_b$. All parameters on this site are proposed and will be fixed at the official launch.

Three things that surprise newcomers

Reading this site cold? These are the things that catch first-timers off guard. None of them are bugs — they're how variable-rate pool lending works on every chain — but it's worth knowing them up front.

1. All Forge rates are variable. Both the borrow rate (what you pay) and the lender APY (what you earn) change every block based on how much of the pool is currently borrowed. The 8% you saw at deposit might be 25% next week, or 3%. There is no way to lock a rate.
2. Forge loans have no fixed term. Direct loans are open-ended — borrow today, repay in 5 minutes or 5 years. Interest accrues silently every block; a loan you forget about keeps growing. (Bootstrap is the exception — bounded by the underlying QNS Staking lock.)
3. Lenders never lock their QOM. Deposit and withdraw any time, subject to free pool liquidity. The "lock + multiplier" pattern you may have seen elsewhere in the QOM ecosystem belongs to ZAP Staking — a separate product on staking.zaphod.one, not Forge.

Three deposit shapes — don't confuse them

Forge composes with QNS Staking, which means three different "deposit your tokens" UX patterns coexist in the same product family. They look similar but have very different lock semantics:

You deposit	Where	Lock?	Earns?
QOM as a Forge lender	Forge pool	No	Variable APY from borrowers' interest payments
ZAP as Direct-loan collateral	Forge pool	No — withdraw whenever you repay your loan	Nothing — collateral just backs the loan
A freshly-minted .qom, via Bootstrap	QNS Staking, with a lien	Yes — until both the loan debt clears AND the staking lock expires	ZAP emissions, auto-routed to repay your loan first

Lenders never lock anything. Direct borrowers never lock anything. The lock-based shape only appears via Bootstrap, and it's because Bootstrap composes with QNS Staking — which is itself a lock product.

Two products, one pool

Direct Loan. The Aave/Compound shape. Deposit ZAP, borrow QOM at up to 40% LTV, pay variable interest along a kinked curve, liquidatable on collateral price drop. Survival distance $\Delta p_{\text{liq}} = 1 - L_0 / L_{\mathrm{LT}}$, with grace and partial-liquidation guardrails.

Bootstrap Loan. No upfront collateral. The protocol mints a fresh .qom, stakes it into QNS Staking with a lien, and routes the position's ZAP emissions back to retire your debt. The closed-form payoff time is

$$t^\* \;=\; \frac{1}{r_b}\,\ln\!\!\left(\dfrac{(1-\alpha)\,\bar e\,\bar p}{(1-\alpha)\,\bar e\,\bar p \;-\; r_b\, D_0}\right),$$

finite whenever the position's cash-flow yield exceeds the borrow APR.

Both loan types draw QOM from the same lender pool. Lenders see a single utilization-driven APY; they don't pick which loan type their QOM funds.

The lender's mental model

Deposit QOM. Earn $R_l(U) = R_b(U) \cdot U \cdot (1 - \phi)$. Withdrawals are first-come, first-served against available liquidity ($D - B$). The kinked rate curve (gentle below 80% utilization, steep above) is what keeps a withdrawal buffer in the pool — a lender on a healthy pool effectively never queues, because high utilization is so unattractive to borrowers that it self-corrects.

Bad debt is absorbed by the protocol reserve, not by you. See Resilience for the full failure-mode analysis and the reserve-growth model.

The borrower's two paths

	Direct Loan	Bootstrap Loan
Upfront capital	ZAP collateral	None
Debt service	From your wallet (or ignore until liquidated)	Auto, from staked-position emissions
Term	Open-ended; repay any time	Bounded by the staking lock duration; closed-form payoff time
Risk surface	ZAP price drawdown	Emission-rate divergence + price drawdown
Best for	"I want leverage on ZAP without selling"	"I want a .qom and the staking yield, no cash up front"
External analog	Aave v3 USDC market	Alchemix v3 (yield-source-deleveraged CDP)

What's different vs. typical pool lending

• Bootstrap path — only possible on QOM. Composes the lending pool with QNS Staking's deterministic emission schedule and registry-level lien hooks. No other chain exposes both primitives together.
• 24-hour grace period before liquidation — a position dipping below $\mathrm{HF} = 1$ opens a top-up window before the keeper auction opens. Single-block oracle noise will not take a solvent position.
• Reserve-funded bad-debt absorption — closer to Aave's Umbrella (first-loss capital absorbs before depositors) than to Morpho Vault V2's pro-rata socialization. Lenders see clean principal+interest at withdrawal.
• Variable rate curve with a steep tail — $s_2 = 60\%$ slope above the kink keeps a withdrawal buffer for lenders.

Read next

• Direct Loan mechanics — health factor derivation, sensitivity grid
• Bootstrap Loan mechanics — the deleveraging ODE, closed-form payoff time, comparison vs Alchemix
• Rate Card — three-tier calibration (conservative / base / aggressive), IRM curve plot, sensitivity grid
• Liquidations — keeper economics, three guardrails formalized
• Resilience — reserve growth model, scenario sensitivity, comparison vs peers