The sovereign onchain central bank.

STANDARD is a closed monetary economy with one currency ($STANDARD), one market (ETH <> $STANDARD pool on Uniswap v4), one signal that affects monetary policy (net ETH flow through that market), and one authority (the central bank, 4,000 lines of immutable code). Bankers (you) hold charters, charters operate bank branches, bank branches earn the currency issued by the central bank.

Capital flowing in loosens policy, increases $STANDARD issuance, and stacks hard reserves (tokenized gold). Capital flowing out tightens policy, triggers buybacks and burns, and prices the exits. Every path through the economy either burns $STANDARD or brings the central bank hard assets. Most do both.

There are six entities. Everything in the system is a relationship between them.

The mental model for charters and branches: a charter is a company, branches are its stores. The company earns through its stores, reinvests earnings to open more, and pays its owner by closing stores, one at a time or all at once. Closing the last store dissolves the company.

The flows between them:

• Traders ⇄ pool. Anyone can buy or sell $STANDARD, no charter needed. Every swap feeds the bank a trading fee in ETH.
• Pool → central bank. The v4 hook on the pool reports net flow: ETH in from buys minus ETH out from sells. This is the bank's only input.
• Central bank → branches. Each epoch the bank issues $STANDARD to all branches, pro rata. A banker's share of the issue equals their share of total branches.
• Bankers → bank. Bankers spend earned $STANDARD on expansion licenses to open more branches. Every token spent this way is burned.
• Bankers → pool. To realize earnings, a banker retires a branch: its share of the balance is liquidated and minted to their wallet, minus the resolution fee.
• New bankers → central bank. New charters are bought at a daily ETH auction. The ETH flows into the same fee engine as trading fees.

• Hard cap: 1,000,000,000 $STANDARD. 18 decimals.
• Genesis: 100,000,000 as protocol-owned liquidity locked into the v4 pool. This is the only pre-mint. The tokens sit single-sided above the launch price with no ceiling, owned by the protocol, and can never be withdrawn; buyers’ ETH becomes the liquidity beneath the price as it discovers.
• Issuance budget: 900,000,000 tokens (the cap minus the genesis liquidity). When cumulative issuance reaches the budget, base issuance stops permanently and the economy runs closed-loop on recycled fees.

$STANDARD is minted on demand. Issuance credits a banker’s balance as a ledger entry; actual tokens are minted only when a banker withdraws. Value leaves the system permanently through two doors: ledger removals that are simply never minted (expansion licenses at 100%, the burned half of resolution and revocation fees) and token burns (open-market buybacks and the protocol position’s token-side fees). Deposits also destroy tokens, but as a conversion, not a removal: the value re-enters the ledger 1:1 and is re-mintable on withdrawal. Supply therefore obeys a single identity at every block:

and everything permanently removed is gone forever, enforced in the token contract itself: every public burn lowers the mintable ceiling maxSupply() by the burned amount, and every ledger removal (a license payment, the burned half of a fee) lowers it by the same amount at the moment it leaves the ledger, even though those tokens were never minted. Only the CentralBank’s deposit-conversion path is exempt. The maximum supply that can ever exist is strictly non-increasing:

where $R(t)$ is cumulative ledger removals. The ceiling never strands value: minted supply plus everything the bank still owes always fits under it.

Circulating supply is a receipt, not a promise.

It equals value withdrawn from the system minus everything the bank has clawed back and burned. One onchain number tells you whether the economy is eating or bleeding.

The pool’s hook counts, per epoch, gross ETH entering from buys and gross ETH leaving from sells. Net flow for epoch $n$ is the difference, and the policy signal aggregates the last two completed epochs:

The issuance rate moves on $signal_n$; fee routing moves on $sign(F_n)$ alone. Properties that matter:

• It is measured at the canonical pool. There is exactly one place ETH enters or leaves this economy through trading, and the bank watches it.
• It is denominated in real capital. ETH, not token amounts, not trade counts. To move the signal you must move actual capital.
• The issuance decision uses the trailing two epochs, so one manipulated hour cannot swing the rate. The fee routing decision uses the current epoch’s sign, so defense reacts fast. Slow lever for issuance, fast lever for fees.

Issuance runs at a base rate of 700,000 $STANDARD per day at launch, scaled by a policy multiplier m. The base rate can be lowered by the owner, taking effect at the next epoch, and can never be raised back. An epoch of $d$ days issues

split pro rata and streamed second by second across the epoch. Balances tick up in real time, and a new branch earns from the moment it opens. A single branch’s daily yield in a system of $N$ branches is $\,700{,}000 \times m / N$. The multiplier itself follows one rule, evaluated every epoch:

The asymmetry is deliberate: cuts are immediate, raises must be earned. Issuance opens at full rate; sustained inflows reach the 1.25× ceiling in twelve days. A sustained exodus drives the rate from ceiling to floor in three weeks, cutting dilution by 84% while it happens. The bank turns defensive faster than it turns generous.

The two regimes. Each epoch is either expansion (net flow positive) or contraction (net flow negative or zero):

A charter is an initially soulbound NFT. Holding one makes you a banker: it is the license to operate a bank and receive issuance.

• Genesis: 1,000 paid Founding Charters. The whitelist mint carries a 0.15 ETH liquidity fee, limit one per wallet. Every unminted charter then enters a public Dutch auction, decaying exponentially for 30 minutes to the whitelist price and remaining there. The public phase is open to everyone, whitelist included, one charter per transaction, up to 3 charters per wallet across both phases. Proceeds escrow until the sale is finalized and fund initial liquidity and the protocol vaults; 0% goes to the team. Founding charters accrue nothing during the sale; they all begin earning together when finalization starts epoch one.
• After genesis: daily Dutch auctions in ETH. Same falling-price mechanism as expansion licenses: the day opens at 3× the last sale, decays toward the reserve floor, and buying is first-come-first-served at the current price, with the charter minted instantly on purchase. The number auctioned per day is policy-controlled and starts at zero; charter supply is never unlimited and never free-flowing. Auction proceeds enter the fee engine like every other ETH flow.
• Lifecycle. A charter lives until its last branch is retired, at which point the NFT burns. The only way back in is buying a new charter at auction. There are no revolving doors.

Every charter opens with its first branch, and can grow to 10 branches maximum. Each branch is one share of every epoch’s issue.

Additional branches require an expansion license, sold at a Dutch auction with 12-hour periods. The owner activates each auction contract once with an opening price; that timestamp anchors its periods and fixed 24-hour purchase windows:

Formally, with $N$ total branches in the system, the floor is two days of one branch’s issuance, and the price decays exponentially from open to floor across each auction:

where $I_{base}$ is the daily base issuance and $P_{start} = 2 \times P_{last}$ after a sold period. If the prior period had no sales, the next opening uses twice the new floor.

The auction starts high and decays until buyers step in, so every license finds its market price; the floor only prevents literal-zero sales. A branch begins earning the second it opens: issuance streams continuously, so its take is exactly proportional to how long it has existed.

This is the engine of the expansion flywheel: the most rational move inside the system, growing your bank, permanently shrinks the float.

Both sales use a falling-price Dutch auction. The price opens high and decays toward a floor. Purchases execute instantly at the current price, first come, first served. There are no bids or escrow. Licenses are paid in $STANDARD and burned; charters are paid in ETH that flows to the fee engine.

The design has two intentional consequences:

• Buyers set the price. Buy early for certainty, or wait for a lower price and risk the period’s supply selling out. The floor prevents literal-zero sales.
• Each auction reprices from demand. A selling period anchors the next opening to its final sale. A period with no sales resets the next opening from the floor. Licenses use a 2× opening multiplier; charters use 3×.

The license auction clock and the per-bank allowance clock serve different purposes. Supply resets each auction. The three-license allowance resets only at the next anchored 24-hour boundary, not at a buyer’s local midnight and not 24 hours after each purchase.

Issuance accrues to a charter’s balance continuously. To take profits, a banker retires branches. Retirement is two things at once: it liquidates the branch’s share of the accrued balance into tokens in your wallet, and it permanently retires the vehicle that was producing the yield.

• Pro rata rule. Retiring one branch of ten liquidates one tenth of the balance. Retiring all ten liquidates everything and burns the charter.
• The released amount is minted to the banker’s wallet, minus the resolution fee.
• Minting has an equal and opposite reaction. Every withdrawal retires the branch that earned it, permanently reducing your share of all future yield. You cannot extract value and keep the vehicle that produced it.

The resolution fee is congestion pricing on the exit door. Let $W$ be tokens withdrawn system-wide over the trailing 7 days and $D$ be everything still held at the bank. Exit pressure and the fee it commands are:

A quadratic curve from a 2% floor to a 60% ceiling, saturating when a tenth of the bank tries to leave in a week:

The fee curve is quadratic between the floor and the ceiling, and your rate locks the moment you commit. Half of every fee is burned. The other half is paid to every banker who stayed.

This inverts the payoff structure of a bank run. In a traditional run, whoever exits first is made whole and whoever waits absorbs the loss, so running first is always correct. Here, heavy exit volume raises the fee on the exiters themselves, and half of what they pay goes to the positions that stayed, so mass exits transfer value from the impatient to the patient. The contract exposes no withdrawal pause or queue at any fee level. The cost of leaving is the only intended control mechanism; independent settlement dependencies can still make a Bank call revert.

A dormant banker siphons issue away from working bankers, so the system removes them:

1. 01

   Report

   A wallet inactive for 30 days can be reported by anyone.

2. 02

   Bounty

   The informant earns a bounty (2% of the dormant balance, capped at 100,000 tokens).

3. 03

   Revocation

   The ghost pays a 70% revocation fee, deliberately worse than the worst-case 60% resolution fee, so going dark is never the cheap way out. Half the fee burns, half pays the bankers still at their desks.

4. 04

   Shutdown

   Their branches are shuttered, their charter burns, and the remaining 30% is sent to their wallet.

Staying active is free: creating a charter, buying a license, depositing, withdrawing, successfully reporting another dormant charter, or calling the zero-cost check-in resets the wallet clock. Sending or receiving a charter does not refresh either wallet, but a transferred charter carries its own timestamp and receives a full 30-day grace period. Lost keys, abandoned wallets, and tourists dilute no one.

The steady-state split. After launch, trading fees and charter auctions route each epoch:

The expansion vault accumulates ETH and purchases hard reserve assets (tokenized gold and comparable assets). Reserves are held by the bank.

The contraction vault buys $STANDARD on the open market and burns everything it buys, executing in small rate-limited steps so defense cannot be baited into one blockable shot. Each hourly tick with vault balance $V$ against pool reserves $R$ spends

bounding buybacks near 5% of pool depth per day at launch settings. Unspent balance rolls forward; the vault can never sell.

POL buybacks use the 15% share to buy $STANDARD and hold it in the Incentives Vault for future distribution. These tokens remain minted; they are not burned. The genesis liquidity positions remain in place and no longer receive new liquidity from the POL share. Their original manager continues collecting LP fees.

Outside liquidity pays the same rates as trading. Third-party positions open once the launch schedule has decayed to its floors, and withdrawing a position’s principal is taxed by the hook as the swap it completes: the ETH leg at the sell rate, routed through the split above, and the $STANDARD leg at the buy rate, burned. Accrued LP fees are never taxed, so a range order is never a cheaper way around the trading fee.

The protocol is non-upgradeable. There are no proxies or code-migration mechanisms. Within that, parameters split into three classes:

1. 01

   Immutable forever

   The 1B cap, the 700k/day base-rate ceiling (the live rate can only ratchet down), the multiplier rule, the resolution fee curve, the 70% revocation fee, the 10% ceiling on manually-set taxes, the launch tax schedule (a pre-committed 90%/90% open that only decays to the 2%/3% floors; the admin can never set rates that high by hand), the vault execution bounds, the founding whitelist price, and the public sale's opening price and 30-minute curve.

2. 02

   Tunable within hard bounds by the owner

   Epoch length (1–7 days), tax rates (≤10%), licenses per auction period (≤2,000), charters per day (≤100), fee splits (team ≤20%), auction floors, windows, and decay half-lives. The founding whitelist window may be changed only before the sale opens; opening freezes its timestamps. At production deployment the broadcast sender/deployer is the sole owner, and its transactions have no protocol-enforced delay, notice period, threshold, or cancellation window. The owner may later transfer selected contracts manually through Ownable2Step to a Safe/multisig, but no recipient is deployed, configured, or assigned by the production launch.

3. 03

   One-way switches

   Charter transferability (off until on, then forever), and permissionless contraction-buyback/POL-buyback execution (owner-cranked until opened to everyone, then forever). ExpansionVault reserve purchases remain owner-only. Both switches only move toward less control.

An optional emergency guardian role may be configured with exactly one power: pausing auctions and vault purchases. It cannot pause withdrawals, cannot touch funds, cannot change parameters, and can be disabled or permanently renounced by the owner.

Charters launch soulbound. A one-way switch enables transfers later, at which point selling a charter becomes a second exit path: the seat moves whole, branches and balance included. A seat sale is an exit with zero sell pressure on $STANDARD; the buyer replaces the seller one for one.

Four structural loops, each following directly from the mechanics above.

STANDARD is an experimental onchain protocol. It is not a bank, holds no customer funds, offers no accounts, and is not a regulated financial institution of any kind. Reserve assets are protocol property and are not redeemable. Nothing here is investment advice. Participate at your own risk.