The idea
Orbi takes the other side of every trade. Pushing him further costs more than bringing him back.
When you buy, you buy from him. When you sell, you sell to him. There is no order book and nobody else to trade against: the pool is the counterparty, and Orbi is what the pool looks like from the outside.
What matters is that he keeps count. Every trade leaves him a little more on one side than the other, and that imbalance is the only thing this protocol stores. He decides nothing. He has no discretion, no opinion, and no ability to refuse. He is a number with a face.
His belly is flow, not reserves
And that distinction is the whole design.
In any automated pool, the reserves and the price are the same fact. If the pool holds a lot of a token, that is only another way of saying the price is low. A market maker who skewed his prices on reserves alone would be repeating the price back at you.
So Orbi's belly is not what the pool holds. It is the net flow of the last few minutes: what was sold to him, minus what was bought from him. Two pools at the same price can have opposite bellies. One has been quietly absorbing sells for ten minutes; the other has been drained by buyers. The price does not tell you which. The belly does. That is what a real market maker watches, and it is called pressure, not level.
He digests one thousandth a block
Toward empty, never past it.
Every block, Orbi's belly shrinks by one thousandth of what the pool holds, moving toward zero and stopping there. Nothing else empties it. On a pool holding $500,000, that is $500 a block. About $3,000 a minute, and roughly a third of the pool over an hour.
Ordinary trading never builds a belly, because ordinary trading is noisy in both directions and the digest eats it. The belly only builds when the flow is genuinely one-sided and genuinely fast. The digest is a straight subtraction, the same amount every block, with no curve and nothing to approximate. You can check it with a calculator.
of what the pool holds. About a third of the pool per hour.
a block, $3,000 a minute, the same amount every block.
The table
Four lines. Two columns. Nothing else.
| Orbi's belly | Feeding him | Emptying him |
|---|---|---|
| under 5% of the pool | $3 | $3 |
| 5% to 15% | $5 | $2 |
| 15% to 40% | $7 | $1 |
| over 40% | $8 | $1 |
Dollars per $1,000 traded. That is the entire fee logic of this protocol, written out. There is no second table, no multiplier, no discount, and no address that can edit it.
The first line matters as much as the last: when the belly is under five percent of the pool, both columns are $3. The symmetry is not a policy we apply in calm markets, it is the top row of the table. The thresholds are a share of the pool, not a dollar amount, so the same table works on a pool holding fifty thousand dollars and one holding fifty million. $8 per $1,000 is the ceiling. There is no belly, no flow, and no combination of anything that produces a higher number.
Feeding and emptying
The protocol does not know what a buy is.
The table has two columns, and which one you pay is decided by one comparison: does your trade push the belly further from zero, or bring it back?
Depending on the minute, buying is the expensive side or the cheap side. The protocol never learns which direction is which, because direction is not stored, only a sign is. What that costs you: the cheap side is cheap precisely when it is unpopular. An asymmetry that only rewards going against the crowd is only useful to people willing to go against the crowd, and most of the time most people are not. Part 10 puts a number on that.
What it costs to push him
Moving the belly means trading, and trading means paying.
The obvious question: can somebody move Orbi's belly to make your trade expensive? They can, and it costs them more than it costs you. To push the belly one tier, they have to trade in that direction, paying the rising fee themselves, on their own volume, at every threshold they cross. And what they inflict on you is capped at the gap between the top row and the bottom.
on real volume, in the direction that fills the belly.
per $1,000 at most, the gap between $3 and $8.
until the digest eats it.
The belly cannot be moved by adding or removing liquidity, by holding tokens, by voting, or by waiting. It moves one way only: by trading, and by paying. What we have not solved: it does work, it is just expensive. Somebody with a reason to make a specific large trade costly can do it. We have bounded the damage; we have not made it impossible.
Where the fee goes
Two places. There is no treasury.
Nothing else. No treasury, no team allocation, no revenue share, no staking, no claim, and no address that can withdraw. The fee moves from the swap into those two places in the same transaction, and the protocol never holds it. That dollar is the entire link between this hook and its own token, and it is also the answer to who benefits when a pool adopts Orbi.
Why a hook
Two things that only happen at the moment of a swap.
A hook is code that a Uniswap v4 pool calls on every swap. This protocol needs two things to happen right there.
- 1The fee has to depend on the direction of the swap.A normal pool has one fee for both sides, decided once. A v4 pool asks its hook, every time, and the hook is the only code that sees which way you are going.
- 2The belly has to be updated in the same transaction as the trade.If somebody had to call a function to report the flow, the number would lag, and a number that lags is a number that can be traded against.
And one thing worth saying plainly about what this hook cannot do: it cannot change what you receive from a swap. It sets the pool's fee, and that is the whole of its power over your trade. It has no permission to touch the amounts, and anybody can read that in the contract: the two settings that would allow it are switched off.
What it cannot do
- 1It cannot refuse a trade.In any direction, in any state.
- 2It cannot change what you receive beyond the fee shown in the table.
- 3It cannot stop a liquidity provider from entering or leaving, nor impose a delay.
- 4It cannot treat two addresses differently.No address enters the computation.
- 5It cannot know what a price is.It never reads the pool's price, and none of its decisions depend on it.
- 6It cannot be paused, upgraded, or reconfigured.
- 7It cannot remove a pool from the board.A pool that adopted the hook adopted it for good.
The flaw
He is fullest when you most want out. This is the part that does not work. Read it before you trade.
A real fall is one-sided flow, fast, for a long time. That is exactly the condition that fills Orbi's belly and pins it at the top of the table. So during a crash, selling is the feeding side, and selling costs $8 per $1,000. The protocol charges the most at the moment people most need to leave.
belly under 5%
belly between 5% and 15%
belly over 40%. Two and a half times more at the worst moment.
The other half
At the same moment, buying costs $1 per $1,000, the cheapest this protocol ever gets. Somebody standing on the other side of that fall is being paid to stand there, and it is being paid by the people running.
Whether that helps depends entirely on whether anybody is willing to stand there. If nobody is, the protocol has simply charged sellers more, and the cheap side sat unused. That is the honest failure mode, and we cannot fix it without deleting the asymmetry that is the entire protocol.
What we refuse to do about it
Cap the fee on sells: that fixes the case and destroys the rule, because the symmetry of the protocol comes from not knowing what a sell is. Narrow the gap: that softens the worst case and shrinks the subsidy by the same amount, the only thing that could bring somebody to the other side. We cannot make the exit cheaper without making the entry less attractive. It is the same gap.
The second flaw
The fee you pay does not depend on you. Two identical trades, two minutes apart, can cost three dollars or eight, and nothing that belongs to you has changed. That is how a congestion toll works, it is deliberate, and it is unpleasant. The only way to know what you will pay is to look at the belly before you sign, which this site shows at all times.