When lending leaves the pool: Midnight's peer-to-peer fixed-rate order book

1delta is a lending aggregator. It reads and normalizes market data and per-user positions from dozens of protocols into one consistent shape, and builds the calldata to act on them - deposit, borrow, repay, one-click leverage, cross-lender migration, spot swaps - behind a single API. Our piece on lending aggregation makes the general case: on-chain credit is fragmented, and the 1delta API collapses thousands of disjoint markets into one queryable surface.

Look closely at that surface and almost every market on it - however different its ABI, its risk model, its idea of a "market" - shares one deep structural assumption. It's a pool. Morpho Midnight is the first lender we've integrated that throws that assumption out. This piece is about why that's a bigger deal than a new rate model, and what it unlocks.

The pool is the hidden default

Pooled lending is so universal it's practically invisible. Suppliers deposit into one shared pot; borrowers draw from it; a utilization curve maps how much of the pot is borrowed to an interest rate that floats every block; and positions are open-ended - you supply and withdraw whenever you like. Aave, Compound, Morpho Blue, Euler, Fluid, Silo, Dolomite - different code, same substrate. The pool is the thing you never think about because everything is one.

Even the exception proves the rule. Lista DAO offers genuinely fixed-rate, fixed-term borrowing - and yet it's still a pool underneath. Lista's fixed-term product is a broker sitting in front of a Moolah market (a Morpho Blue fork): the term and the rate are locked, but the liquidity and collateral still live in a shared pool, and the supply side still earns the floating pool rate. Fixed-rate, pooled-liquidity. The pool didn't go anywhere; a fixed-term wrapper was bolted on top of it.

That's the standard Midnight breaks from.

Midnight leaves the pool

Morpho Midnight is a true peer-to-peer order book for fixed-rate, fixed-maturity, zero-coupon loans. There is no shared pot and no utilization curve.

Peer-to-peer. Lenders and borrowers post signed limit offers at a rate they choose. A taker fills an offer directly - one counterparty matched to another, not a deposit into a communal pool. Rates aren't computed by a curve; they're discovered by the book, the way a real fixed-income desk quotes.
Zero-coupon. A loan is a claim on a fixed face value at a fixed date: you receive assets now and owe (or are owed) a set amount at maturity. The discount between the two is the interest - so the rate is baked into the price of the offer, not accrued block by block.
Fixed maturity. Every market has one hard calendar maturity shared by everyone in it. Time-to-maturity is a first-class term, not an afterthought.

Because there's no pool, the familiar pool concepts simply don't apply. "Total deposits" isn't a pool balance - it's the depth of resting offers. Utilization is undefined. And a position isn't open-ended - it has an end date. This isn't a new curve on an old pool; it's a different market structure.

Why leaving the pool matters

You choose your side of the price. In a pool you're always a price-taker of the curve - the rate is whatever utilization makes it. In an order book you get both roles: take the best offer resting on the book right now, or make your own limit offer at a rate you set and wait to be filled. Being able to be the price-maker - to say "I'll lend at 4% and not a basis point less" - is something a pool can't give you.

It's real fixed income. Fixed rate plus fixed maturity is a bond-like instrument: a predictable cost of capital for the borrower and a predictable, locked yield for the lender, with no rate surprises mid-position. That's exactly the primitive that's hard to build on top of a floating pool.

It also comes with new responsibilities. There's no pool to redeem from, so exiting early means trading your position back into the book rather than withdrawing from a pot - a lender sells their claim, a borrower buys their debt back. And because the loan has a hard maturity, a borrower has a deadline: repay by the maturity date or the loan falls into default and can be liquidated regardless of collateral health - being past-due is itself the trigger, even on a comfortably-collateralised position.

Repaying a Midnight loan, exactly

Repayment is the one place the order-book model quietly changes what a familiar verb means, so it's worth spelling out. A pooled loan accrues interest every block, so pooled protocols happily let you over-pay and refund the dust, or accept a repay-max sentinel and sort it out for you. A Midnight loan is the mirror image, and getting it right is mechanical:

Read your exact debt. It's a static face value in the loan token - it does not drift between blocks - so a single read of your position is the exact figure to clear, not a moving target. The two fees a Midnight market carries (a continuous fee and a settlement fee) are netted from the lender's proceeds, never added to your debt, so the face value is the whole bill.
Approve exactly that amount of the loan token to the Midnight contract.
Repay exactly the debt - the bounds are strict on both sides:
- Over-repaying reverts. There is no over-repay buffer and no refund; sending more than you owe fails the transaction on-chain. You can't pad the amount "to be safe" - which is precisely the habit a pooled protocol trains.
- Under-repaying leaves dust. Whatever you don't clear stays open as a live position, and an open position past maturity is still liquidatable. Dust is not harmless here.
Then reclaim your collateral. Collateral can't be withdrawn while any debt remains, so a full close is two legs: repay the face value, then withdraw the collateral. There's no early-repayment penalty for closing before maturity (unlike Lista) - you owe the same face value whether you repay today or at the maturity date.

So the compliant repay is simply exactly the debt, before maturity, then withdraw the collateral - and because that amount never drifts, it's knowable up front. The 1delta repay endpoint does the sizing for you: point it at the position's current debt and it builds the approve-and-repay at the exact face value, so you never have to reason about the revert bounds by hand.

When the repay is one leg of a bigger move. You rarely hold the loan token in the two flows where you'd repay without simply paying it off: unwinding a leveraged (looped) position, or rolling the debt into a new term or another lender. There a flash loan, routed through the 1delta Composer, supplies the exact face value, repays the Midnight loan, and frees the collateral - which then either repays the flash (a close) or backs a fresh borrow (a roll / migration), all in one atomic transaction. The exact-debt rule gets stricter here, not looser: an over-repay reverts the whole bundle, and any dust left behind blocks the collateral-withdrawal leg the rest of the bundle depends on. The loop-close and migrate builders size that repay to the face value for you, so the atomic unwind clears in a single signature.

Still one surface

Here's the part that makes this an aggregator story rather than a one-off integration: despite being structurally unlike every pool around it, a Midnight market lands in the same normalized shape as everything else. It carries a fixed-term rate card you can list, a fixed APR you can sort right next to variable ones, and positions you can render in a portfolio - the order-book machinery stays behind the mapping. The only thing that surfaces the difference is how you transact: an order book has a take side (fill existing offers) and a make side (post your own), and the API exposes both.

For the mechanics of how we reconcile Midnight's order book with Lista's broker into one comparable term-card model - the rate math, the two data pipelines, the position shape - see the deep dive: How we unify fixed-term lending.

Takeaway

The pool is the silent default of on-chain lending - even most "fixed" lending is a fixed wrapper over a floating pool. Morpho Midnight is the first market in the 1delta index to leave the pool behind entirely: peer-to-peer, fixed-rate, fixed-expiry, price discovered by an order book. It's the most drastic deviation from the standard we aggregate - and we make it comparable to everything that didn't move.

Where to go next

How we unify fixed-term lending - the technical deep dive on Lista vs Midnight and the shared term-card model.
Aggregating on-chain lending - why fragmented lending data needs one surface.
API reference - endpoints, the take/make model (Actions › Midnight), and x-api-key setup.