- fees
- creators
- launches
- mechanics
A launch costs a dollar
The creation fee used to be a percentage of what you were trying to raise, which charged the most ambitious launches the most simply to exist. It is a flat dollar now — the same for a two-ETH raise and a hundred-ETH one — because the platform's money comes from trading, not from listing.
A percentage creation fee is a tax on ambition, and the arithmetic is unforgiving about it. At a tenth of a percent, a creator aiming at a hundred-ETH raise pays fifty times what a creator aiming at two ETH pays — for the same contracts, the same deployment, the same everything. Nothing about a larger target costs the platform more to create. The fee scales anyway, because that is what a percentage does.
It also lands before anything has happened. The creator pays it at the moment of creation, on a raise that has not started, on a target that may never be met.
The fee is $1 flat now. A hundred-ETH launch and a two-ETH launch cost the same dollar, and a prediction market costs the same dollar as both.
Built, not yet live
The flat fee is written into the contracts and is not deployed. A launch created today is priced by whatever model is actually running. Read the fee off the factory before you sign anything, which is good practice regardless.
The signature is the argument
The function that quotes the fee takes no target. That is §1 of the change in one line: there is no argument to pass because there is nothing about your launch that changes the price.
It would have been easy to keep the argument and ignore it — but a function that accepted a target and returned a figure independent of it would be a lie every caller went on passing, for ever. The argument was removed rather than ignored.
Why a flat fee is not the platform being generous
This is the part worth being precise about, because "we made it cheaper" is marketing and the structure underneath it is not.
The creation fee is the small end of what a launch is worth to the platform. The trade fee is the business. Every trade on the curve pays a fee, and after graduation the pool keeps charging one — so a token that trades for years goes on paying for years. Measured against that, what the platform collects at the moment of creation is a rounding item.
Which means the percentage fee was defending the small end at the cost of the large one. It priced out exactly the creators whose launches would have been worth the most, to protect the part of the revenue that matters least.
The creator is on the same side of that fee
A creator keeps a share of every trade fee on their own token, for as long as it trades — including after graduation, without the token having any fee logic in it at all. Where a trade fee actually goes has each leg and who decides it. The platform and the creator are paid by the same event, which is what "we earn when you earn" has to mean if it means anything.
Turning a dollar into wei without becoming somebody else's outage
A dollar figure needs a price, and a price means an oracle sitting in the path of every launch. An oracle in that path is a way for somebody else's outage to stop the platform taking new projects, so the resolution order is built to fail in the right direction:
| Condition | What happens |
|---|---|
| A posted price, refreshed roughly every fifteen minutes, is fresh | That price is used. Nothing else is read |
| The posted price is stale or absent | Fall through to the live aggregator behind it |
| The live feed answers within its bound | That price is used, and the read says which source answered |
| Neither can answer | Revert, naming both feeds, both last-update times, both bounds and the current time |
The bounds are not left to judgement. The posted cache is capped at six hours by a constant — not a setting, a constant, so nobody can widen it past the point where a fee stops being a fee. The live feed's bound is capped at thirty hours, which is the outer edge of what this platform is willing to call a price anywhere.
A fifteen-minute-stale price applied to one dollar is worth cents. A launch that cannot happen is worth considerably more than that, which is why the design prefers a slightly stale dollar to a correct outage.
Quote it immediately before you create
The figure moves with the ETH price and steps each time a new price is posted, so a quote is good for minutes rather than hours. Attaching too little reverts with an error naming the exact shortfall. Attaching too much costs nothing at all — everything above the fee becomes your opening buy on your own curve.
A ceiling on the fee, in the contract
The dollar figure is settable, which is what lets the platform run a free-creation period or correct a mistake. Left there, "settable" would mean the wall could quietly come back.
It cannot go above $10, and that is a constant in the contract rather than a policy. Whatever happens to the setting, the fee has a ceiling that no transaction can lift.
A fee of zero is a special case in the other direction, and a useful one: a zero fee never touches the price oracle at all. So a free-creation period is one transaction, and it doubles as the escape hatch if the price feeds ever go dark together.
What the flat fee quietly fixes
A free creation is worth the same to everyone. Under a percentage, a free-creation promotion was worth a few dollars to a small creator and hundreds to somebody aiming at a thousand-ETH target — the same prize, wildly different value, and the obvious thing to do with it was to point it at the largest target you could type. At a flat dollar the prize is worth exactly a dollar, every time, to everybody.
The graduation ladder gets simpler. The percentage fee had to divide without leaving dust, which meant every rung on the graduation ladder had to be a whole multiple of a number derived from the fee — and changing the fee re-validated the whole ladder. A flat fee performs no division, so the rule and its per-rung check are both gone. One fewer invariant that could have been violated quietly.
The same dollar on the market side, and there it buys something
A prediction market costs the same $1, and on that side the fee is not only revenue — half of it is posted straight into the market as an opening stake on each side, so the market is born with a counterparty on both. Twenty-five percent to YES, twenty-five percent to NO, and the market can never end up in the empty-pool case where everybody is refunded and nothing settles.
Every market opens with money on both sides has that in full, including what happens to the platform's seed once the market settles.
Reading it yourself
launchFeeWei()on the factory — what a launch costs right now, in wei. It takes no argument; if you find yourself wanting to pass a target, that is the point.launchFeeUsd— the dollar figure, at eighteen decimals.1e18is one dollar.MAX_LAUNCH_FEE_USD— the ceiling. A constant.creationFeeOf(currency)on the settlement registry — the same question for a prediction market in a given currency.- The oracle exposes both feeds' readings, both ages and which one would be used, so a disagreement between the cache and the live feed is visible rather than resolved silently in the platform's favour.
The fees and payouts page carries the current figures and the exact call signatures. Where it and this article disagree, that page is newer — and where the chain and either of them disagree, the chain is correct and both are wrong.