Stonkyard is a protocol that turns trading fees into yield for the people holding its products.

Every trade on chain pays a fee to somebody. On almost every venue that somebody is the team, and holders get a token and a roadmap. We route it the other way. Fees are split by a contract in the same transaction that earns them, and spent on assets that go straight to holders.

Two products do that today.

The launcher. Coins launched here convert their trading fees into tokenised stock for their holders. Every trade pays —, and — of that is set aside to buy the coin’s reward stock on the open market and credit it to holders pro rata. A coin can pay a single stock or rotate through a basket of them.

Custom pairs. A coin can launch priced against a stock instead of ETH. Its fees then arrive in that stock already, so there is nothing to convert: — goes out to holders in the pair asset itself.

The vaults. A vault is an NFT with a custody account attached. — of every launcher fee, most of what every new vault pays to mint, and every secondary royalty go into a shared pot. When the pot fills, anyone can run a round: it buys one stock and credits every vault the same amount. You collect whenever you want, and the stock waits in its custody account until you do.

Different products, same rule. Revenue arrives, and holders are what it gets spent on.

Every fee this protocol earns passes through one of these, and all of them are public. Nothing below is a claim you have to take on trust — open any of them and the transactions are the record.

Every contract in that table is owned by the timelock. Nobody holds a key that can move the pot or the pools, and every change waits — in public before it can execute. The governance page shows the delay, the operations in the queue, and who can propose. The rotation and the rest of the live state are at the bottom of this page, read from the chain.

Vaults and the launcher are what we run today. Neither of them is what Stonkyard is.

Nothing on chain stays busy forever. Whatever has volume this month will have less in six, and a protocol built around one product has an expiry date whether it admits it or not. Plenty of them are, and they spend the whole decline saying volume is temporarily soft.

So we did not build it that way. Stonkyard is the machine that turns revenue into holder rewards. A product is just whatever is feeding it at the time. When one slows down we build the next one and plug it into the same router, the same split, and the same contracts. We do not sit and wait for the first one to come back.

The vaults came first and still run, still buying stock every time the pot fills. The launcher came second and is where the volume is now. Whatever is third works the same way, because that part was built to outlast any single product.

That is why the revenue page is the one worth reading. It measures the machine, and the machine is what carries over.

$STONKYARD is our token. It is launched on Pons, not here. It is not a governance badge and it is not a fee discount. It does four things.

Revenue buys it and burns it. A fixed — of every launcher fee goes to a contract that buys $STONKYARD on the open market and burns it. Anyone can trigger the buyback once its balance clears — ETH and — has passed since the last one, and earns — of it for doing so. The buy and the burn happen in one transaction, so a purchase can never sit unburned.

It is what a vault costs. Minting a vault burns — STONKYARD. No purchase, the tokens are just gone. That destroys supply the buybacks never have to, and it costs the protocol nothing, because the person minting paid for it.

Staking it earns ETH. — of every fee is streamed to whoever has staked $STONKYARD in the dividend vault, over —, in proportion to how much is staked and for how long.

Its own trading fees can pay the vaults, and buy itself back. A Pons launch can name a fee recipient. If that is set to our fee splitter, $STONKYARD’s own creator fees are split — vault pot · — buybacks and burns. The pot’s share is spent on stock for vault holders; the other share buys $STONKYARD off the market and burns it. Nothing depends on this being set: it is optional plumbing, and the rest of the protocol runs the same either way.

All of it is on the revenue page, listed separately. Spending revenue and destroying supply are two different claims and they should not be added together.

A launch is a coin on a bonding curve that graduates into a permanently locked Uniswap V3 pool. The curve sells 793,100,000 of the coin’s 1,000,000,000 tokens and raises — doing it. When it sells out, the ETH it raised and the remaining 206,900,000 tokens are minted as a full-range position in the coin’s pool (the 1% fee tier), at the price the curve ended on. The position belongs to the liquidity locker, a contract with no function that removes liquidity, no owner and no upgrade path: the liquidity can never be withdrawn, by the creator or by the protocol.

The pool itself exists from the moment of launch, created by the launchpad and priced at the graduation price in advance, and the coin refuses every transfer into it until the curve sells out, so nobody can seed it or trade it early. Trading after graduation still goes through this site’s buy and sell, which route through Uniswap and keep every trade on the coin’s book; swaps made directly on Uniswap are real trades too, they just do not appear there.

The one thing that is set at the moment of creation and can never be changed: where the fees go. Every trade on the curve pays — to the fee router, and the router splits it in the same transaction. After graduation the launchpad takes nothing; the pool’s 1% swap fee accrues to the locked position instead, and anyone can harvest it into the same router and the same split. That routing is what funds everything below.

You are the coin’s creator and it is credited to you. What you give up is the fee stream, and what your holders get for it is below. A launch can also carry an image of up to 24,000 bytes, written into the chain itself rather than pinned somewhere that can go missing; we call those coins inscribed.

Every trade on a launched coin’s curve pays a — fee. The launchpad hands it to the fee router in the same transaction, and the router splits it there and then, so nothing sits unsplit and no worker has to claim it. Once a coin has graduated its trades pay the pool’s 1% LP fee to the locked position instead, and a permissionless harvest collects it: the ETH side goes through the router exactly like a curve fee, the coin side is burned.

Everything below the holders’ share comes out of the protocol’s side, never out of theirs. The buyback and the dividend are funded by the protocol taking less.

A coin paired against a stock earns that stock rather than ETH. Its holders are paid — in the asset the coin already trades against, in kind and in the same transaction. The rest is held by the router until anyone calls convert, which swaps it to ETH at an oracle-bounded price and splits it four ways, because the vault pot holds ETH and the buyback exists to spend it.

The buyback buys and burns in one transaction, bounded by the Chainlink price, so there is no version of it where the ETH is spent and the tokens are not destroyed. Every buy and every burn is recorded as it happens.

— of every fee is paid to whoever stakes $STONKYARD in the dividend vault. In ETH.

Rather than a snapshot on a clock, every ETH that arrives is streamed over —. A stake earns in proportion to how much is staked and how long it is there, so a balance that shows up for one block around a distribution earns nothing, and one that sits for the whole stream earns its full share.

This is the opposite motion to the buyback on the same token. One takes $STONKYARD off the market and burns it; the other hands ETH to the people already holding it. Both are funded out of what the protocol would otherwise keep.

The holders’ — accumulates as ETH against the coin. Once it reaches — ETH anyone can run a distribution: it swaps up to — into the coin’s next reward stock and credits every holder their share. Whoever runs it earns — of the amount as a bounty.

The threshold exists because a swap gives up slippage. Below it those costs are a real fraction of the trade, so a coin waits rather than paying them repeatedly on small amounts. A coin under the line is not being skipped. Its balance sits and grows until it crosses.

Shares are exact. The coin keeps one accumulator per reward stock and adjusts every holder’s entitlement on every transfer, so what you are owed follows what you held while the rewards came in, not a snapshot somebody took. There is no minimum and no rounding you out. A wallet holding a thousandth of a percent is credited a thousandth of a percent.

A distribution never depends on a single stock. One the oracle cannot price, the venue cannot fill or the registry has banned is skipped for that round, and if nothing in the basket can be bought the ETH is wrapped and credited as WETH instead. Holders are always paid.

Credits are cumulative. Nothing expires, nothing is forfeited for not paying attention, and a wallet that never looks accrues exactly as much as one that checks daily.

Rewards are yours to pull. A distribution credits each holder their share; you collect it from the coin’s page, one stock at a time or all at once, and it lands in your wallet in the same transaction. Anyone can also claim on your behalf, and the stock only ever goes to you.

There are no token accounts to open and no rent on Robinhood Chain. Being paid in a stock you have never held costs nothing extra, and nobody has to pay it for you.

Each coin holds its own reward reserve. Hold ten coins that all pay the same stock and each of them credits you separately, so there are ten balances to claim, not one.

Payouts are listed as they happen on the rewards page, one row per transaction, each linking to the transfer on chain.

A launch picks what its holders are paid in: a tokenised equity, a pre-IPO name, or a basket of up to five that rotates. Any asset the reward registry allows is available, and anyone can register a new one that passes an automated audit, on chain, in one transaction.

The audit refuses anything that cannot actually be paid out:

Governance can ban an asset that turns out to be unpayable. A ban stops new launches from choosing it and makes existing coins skip it when they distribute; it never touches what is already credited. Registrations and bans are events, so a token that passed today and rots later leaves a trail rather than a stale verdict.

A vault is an NFT that owns a custody account. The account fills with tokenised stock over time, and whoever holds the NFT owns whatever is in it. It is the second product here. The launcher is the first.

Minting burns — STONKYARD, pays a — ETH surcharge, and issues an ERC-721 with a custody account at an address derived from its id. The burn happens inside the mint call, before the NFT exists. Transactions are all-or-nothing, so there is no version of this where somebody gets a vault without the supply going down.

There is nothing to activate. ERC-20 balances need no accounts, so a vault is live the moment it is minted, for every stock in the rotation and for any stock added later. — of — have been minted.

The custody account is what makes it a vault rather than a picture: it holds real stock, it is owned by the NFT rather than by you, and it travels with the NFT when you sell. It exposes only withdrawals, never an approval, so a seller cannot leave behind an allowance that drains it after the sale.

When the pot crosses — ETH, anyone can run a round. It converts up to — into the next stock in the rotation and credits every vault equally. The rotation decides which:

A stock the oracle cannot price or the venue cannot fill is skipped for that round, and governance can retire a stock from the rotation without touching what is already owed in it. Retired stocks stay settle-able and claimable forever. Whoever runs the round earns — of it as a bounty, which is what makes it worth anyone’s while to press the button; the operations page lists everything that can be run.

Crediting is a single number. The contract tracks how much of each stock has ever been paid per vault, and your vault tracks where it last collected. What you are owed is the difference. So a round writes one number no matter how many vaults exist, and claiming yours does not move anybody else’s.

Selling the vault hands over everything: the NFT, its custody account, the stock inside it, and anything credited but not yet collected. There is no exit function and no protocol marketplace. The custody account reads its owner from the NFT on every call, so ownership follows the NFT wherever it goes, with no migration and nothing to re-link.

A — royalty is set on the collection and paid to the pot, so every secondary sale funds the next round. It is a marketplace convention rather than something the chain enforces, so it is a behaviour, not a guarantee.

Every contract is owned by a timelock. A change is a public proposal that waits — before it can execute, and anyone can watch it in the queue on the governance page. Ownership cannot be renounced, because an ownerless protocol could never retire a dead stock or rotate a venue.

Some things cannot be changed by anyone. The fee split. The maximum number of vaults and the cap on free grants. The credits a vault or a holder is already owed. There is no function that withdraws the pot or a graduated coin’s Uniswap position, and a new oracle must price every active stock within a factor of two of the old one before it can replace it, so a swap cannot be silently repriced.

Every address this runs on, live from the chain.