Docs

You pay Ξ0.002 plus gas. One transaction deploys a token with a fixed supply of 1,000,000,000, opens a Uniswap v4 market for it against native ETH, and locks the liquidity permanently. Trading starts immediately. You earn 70% of every trading fee that market ever generates.

Most launchpads run a synthetic curve, then “graduate” to a real exchange later. That migration is a discrete event, which makes it something to be front-run, and until it happens the market is not a market — it is a queue.

stfusend skips it. The entire supply goes into a single Uniswap v4 position spanning a ~1000× price range, priced at the top of that range with zero ETH on the other side. Buyers walk the price down through the range exactly the way a curve would, except there is nothing synthetic about it: real depth, real routing, a real chart, from the first block.

The burn’s half of the launch fee is capped in the contract — an owner can raise the burn share but can never push it below half. Your 70% is fixed at the moment you launch, so a later change to the protocol’s split cannot reach backwards into your market.

For two blocks after launch, buying is restricted. In the launch block only the creator can buy. In the following block anyone can, but no wallet may hold more than 5% of supply and no single purchase may exceed 5.5%. After that the restrictions expire permanently and cannot be re-armed by anyone.

The holding cap applies to ordinary transfers as well as purchases, so a buyer cannot route around it by shuffling tokens between wallets. This makes the cap meaningful, but it is not absolute: a determined sniper with many funded addresses can still accumulate. It raises the cost, it does not eliminate the behaviour.

One consequence worth knowing before you launch: the cap binds the creator too. At the opening price, the largest initial buy the contract will accept is roughly Ξ0.18. Send more and the launch transaction reverts rather than quietly handing you a dominant position.

A Uniswap position is only active while the current price sits inside its range, and a stfusend launch opens exactly at its top edge. So the pool reports zero active liquidity until the first buy crosses that boundary and switches the whole range on. This is correct behaviour, not a fault — but charts and aggregators may show the market as empty until somebody trades.

Guaranteed by the contracts: the liquidity position is held by a contract with no withdraw, burn, or transfer path, so nobody can pull it — including us. Supply is fixed with no mint function, no owner, and no pause. The factory ends every launch holding zero tokens, so creators receive no free allocation.

Not guaranteed by anything: that a token holds value, that anyone keeps trading it, that the creator does not sell, or that the price does not go to zero. Removing specific failure modes is not the same as making something safe. Read the risk disclosure.

The protocol owner can change the treasury and burner addresses, the launch fee (capped at Ξ0.05), the launch-fee split (the burn always keeps at least half), the creator split for future launches (never below 50%), the burner’s TWAP and slippage parameters within fixed bounds, and can pause new launches.

The owner cannot withdraw locked liquidity, mint tokens, take a creator’s accrued fees, or extract revenue from the burner. Pausing stops new launches and does not touch existing markets.

One billion, minted once, no mint function, no owner. Protocol revenue buys it on the open market and sends it to the burn address. Anyone can trigger a burn; the contract derives its own minimum output from a 30-minute time-weighted average price, so the call cannot be sandwiched. The burner has no withdraw function — once revenue arrives, buying and burning is the only way it leaves.

$STFU is not a share, a claim on revenue, or a promise. Burning reduces supply; it guarantees nothing about price.

These contracts are not audited. They are covered by 52 tests running against the real Uniswap v4 singleton on a mainnet fork, including adversarial cases and fuzzed money-path invariants. Testing is evidence, not proof. Treat the code as unaudited software holding real funds, because that is what it is.