Protocol reference

The complete mechanics.

A launchpad on Base where every token starts on a bonding curve, graduates into a Uniswap v4 pool with permanently locked liquidity, and pays its trading fees in STITION instead of in the asset it was traded for. You can launch against any token on Base by pasting its contract address.

What this is

Anyone can create a token here for a flat fee. It opens for trading in the same transaction, priced by a formula rather than by a market maker, and it needs no seed liquidity from the creator. Once enough of the quote asset has come in, the launch graduates: the curve hands its reserves to a real Uniswap v4 pool and stops trading.

Every trade pays a fee, 1% by default and up to 3% if the creator chose so at launch, before and after graduation. That fee never stays in the asset it was paid in. It is swapped into STITION, the platform token, and split between the creator and the protocol. The protocol's share is mostly burned. The creator's share is locked for as long as they committed to at launch, and the length of that commitment is what decides how large their share is.

Nothing about a launch's economics can be changed after it is created. The supply, the fee, the quote asset, the creator's lock and their share of the fee are all frozen in the launch transaction.

Launching a token

Launching costs 0.0005 ETH and one transaction. The creator picks a name, a ticker and a logo, and then five things that matter economically:

• The quote asset. Any token on Base: paste its contract address, and Hyperstition checks that it has a swap route and normal transfers before launch. Or pick ETH, STITION or one of the tokenized stocks. This is what the token is priced in on the curve and what its Uniswap pool will be paired against. It cannot be changed later.
• The graduation target, for ETH launches: 4.2 / 2 / 1 / 0.5 / 0.1 ETH, 4.2 ETH by default. Every target keeps the same curve shape and the same supply split; it only sets how much ETH the curve raises before it graduates, so the FDV at graduation runs from 0.49 to 20.58 ETH. Launches quoted in any other token use that token's own scaled curve instead.
• The reward lock, 7 to 365 days. A longer lock earns a larger share of the trade fee, up to 98% of it at a year. See the lockup.
• The trade fee, 1% by default and anything up to 3%. It is charged on the curve and in the pool alike and split with the protocol on the schedule below, so a higher fee means more for the creator and more for the protocol in the same proportion.
• An optional creator tax, up to 10%, charged on top of the trade fee and kept by the creator in full rather than split with the protocol. It makes the token more expensive to trade, so most launches leave it at zero.

A creator can also buy in the same transaction that creates the token, which is the only way to be first without racing anyone. That buy is exempt from the snipe tax below.

The launch window

For the first 15 seconds a buy pays an extra snipe tax that starts near 99% of the amount spent and halves 15 times across the window, reaching zero when the window closes. Combined with the ordinary fees it is bounded so a buyer always receives at least 1% of what they spent, which in practice means an effective ceiling of 98%.

The tax is not a fine that disappears. It joins the ordinary fee bucket, so it becomes STITION and splits between the creator and the protocol like any other fee. The creator, their fee recipient and up to 32 wallets they name at launch are exempt, so a team's own opening buys clear at the untaxed price while bots in the same seconds do not.

How the curve works

The curve is a constant-product market maker, the same arithmetic as a Uniswap pool, with one difference: part of its quote reserve is phantom. The curve prices against 1.68 ETH it does not hold. That is what lets a launch open at a sane price with no seed liquidity.

With the default configuration a launch opens at 1.68e-9 ETH per token, so the first ETH of buying is spread across hundreds of millions of tokens. Buys and sells both work the whole time the curve is live, at the same formula and the same fee, so an early buyer is never locked in.

The curve holds the full supply of one billion tokens but will only ever sell 714.3 million of them, which is 4.2 / (1.68 + 4.2) of the total on the default target, and the same share on every smaller ETH target, since each keeps the phantom at 0.4 times its threshold. The rest is its reserved floor, and reaching that floor is exactly what graduation is. The buy that would cross the line is clamped to the amount that lands on it, and the unused part of the payment is refunded in the same transaction, so nobody overpays for the privilege of being last.

Tokens or quote asset sent to the curve by hand do not count. The curve prices against reserves it tracks itself, not against its balance, so no one can move a launch's price or delay its graduation by donating to it.

Graduation

Graduation unlocks once the curve has taken in its target of real quote asset, 4.2 ETH on the default tier. Anyone can trigger it; in practice our keeper does it within a block or two, and the buyer who crosses the threshold can also settle it in their own transaction.

1. 01Curve fullThe reserved token floor is reached and the target quote (4.2 ETH by default) has come in.
2. 02Fees sweptOutstanding fees leave the curve for the converter before the reserves move.
3. 03Pool createdA Uniswap v4 pool opens at the curve's closing price, seeded from the handed-over reserves.
4. 04Trading movesThe curve closes for good and the site trades the pool instead.

The phantom reserve cannot be handed to a real pool, so the token side of the seed is scaled down to match the quote the curve actually holds. That keeps the pool's opening price identical to the curve's closing price.

The two halves go to two different places. What differs is who can ever reach the tokens:

• The liquidity position holds 204.1 million tokens and the quote beside them, and it goes into the locker. The position is locked forever, so nobody can pull that liquidity, not us and not the creator. The tokens inside it are still buyable, because locking a position stops withdrawals and not trades. The locker has no withdraw function, no owner escape and no upgrade path.
• The 81.6 million the scaling leaves over go, through the locker, into the staking vault. From the moment of graduation they vest linearly over 365 days to whoever has STITION staked, in this launch's own token. Nobody can take them out early, and nobody but stakers can take them out at all.

The position is a full-range Uniswap v4 position, minted straight into the locker rather than to us. Fees the pool earns are the hook's business rather than the position's, so locking it costs nobody anything.

Fees

One fee, 1% of the traded amount by default (a creator may set it as high as 3% at launch), charged the same way before and after graduation. On the curve it is taken on the quote leg of every buy and sell. In the graduated pool the hook takes it inside the swap. There is no separate listing fee, graduation fee or withdrawal fee.

The fee splits into two legs. The creator's leg is 70% to 98% of it, decided by the lock they chose. The protocol's leg is whatever is left. Both legs are swapped into STITION; only then do they part ways.

A creator tax, if the creator set one, rides on top of the trade fee and skips the split: all of it is the creator's. It is still converted into STITION and still lands in their reward lock, so a creator tax buys STITION exactly like the ordinary fee does. The two together are capped at 20%.

What that means in numbers

At a one-week lock, the least a creator can commit to and therefore the protocol's largest share, a million dollars of trading volume on a graduated launch works out as:

The whole $10,000 is buy pressure on STITION whatever the split, because both legs are bought before either is distributed. A launch on a one-year lock buys the same $10,000 of STITION, keeps $9,800 for the creator and still burns $180 of it.

Everything becomes STITION

This is the part that makes the platform token more than a badge. Fees are not collected in ETH and then spent on STITION later at someone's discretion. The conversion is part of the trade.

In a graduated pool the hook takes the fee and swaps it into STITION inside the same swap, along the route the converter holds for that asset. An ETH-quoted launch is one hop; a stock-quoted launch is the stock to ETH, then ETH to STITION. A STITION-quoted launch needs no route at all, because its fee is already STITION.

On the curve there is no pool to swap through, so fees accumulate on the curve and our keeper sweeps them. The sweep converts them the same way. Graduation always sweeps first, so a curve never carries fees into its pool.

When the swap cannot happen

Converting inside someone else's trade must never make that trade fail, and a large fee must never push the STITION price around. So each hop is bounded to 3% of price impact, and the conversion falls through a ladder:

1. Convert. The route exists and the amount fits inside the impact bound. The fee becomes STITION in this transaction.
2. Defer. There is no route yet, or the amount is too large for the bound. The converter records what it holds and for whom, and the keeper converts it later in pieces that fit. The amount is owed, not lost, and it is visible on chain.
3. Never revert the trade. Whatever happens to the fee, the trader's swap settles.

The burn happens once the protocol's STITION reaches the treasury, which counts everything that arrived since the last settlement and destroys 90% of it. Anyone can call that settlement; the keeper does it on a schedule.

The STITION token

STITION is what connects the fee economy of individual launches to a shared asset. Creator rewards are denominated in it, the protocol settles its share in it, and the burn allocation permanently removes it from supply. Its roles in V1 are these and no others:

Supply and the first launch

STITION is intended to be the first token launched through Hyperstition, on the default ETH-quoted configuration described above. So it starts at one billion tokens, the curve sells 714.3 million, the pool is seeded with 204.1 million, and 81.6 million go into the staking vault, where they vest to STITION stakers over 365 days. So STITION's own remainder is paid in STITION, to whoever stakes STITION. The pool's liquidity position is locked forever, which is a separate thing from those 81.6 million.

Two things about that are easy to misread. The 81.6 million in the vault are not out of supply: they count in total supply from day one, and they enter circulation gradually as stakers claim them, so wherever we report circulating supply we subtract only the part that has not vested yet. The 204.1 million pool tokens are different again: they are supply and they are reachable, because locking a liquidity position prevents withdrawing that position, not buying from it. STITION burns reduce live total supply as a third, separate event. There is one timing exception for STITION itself: its stream starts a few hours after launch rather than at graduation, so that stakers exist before the first slice vests.

STITION's own creator fees

For the STITION launch itself, the intended creator-fee recipient is the project treasury, which receives that launch's creator entitlement under the same reward lock as anyone else. That is in addition to the protocol allocation it books separately.

For an independent creator's launch the creator entitlement is theirs. It does not become project revenue merely because the reward is paid in STITION.

At the seven-day tier and the 90/5/5 split, a million dollars of ordinary volume on STITION's own launch allocates as follows, before conversion costs:

So the treasury's combined take in that example is $7,150. That does not merge the two accounting records and it does not remove the creator lock. A higher creator share changes all four numbers, and the burn/treasury/team policy stays subject to the owner powers described below.

Treasury capital is intended to support liquidity and to acquire selected assets in the ecosystem, which V2 would make a native protocol function. Treasury-held STITION is not burned STITION: capital put into a pool or exchanged for another asset can return to circulation.

STITION's V1 utility does not include a right to redeem treasury assets, to vote on spending, or to receive tokenized-stock dividends. Any such right would need its own specification. The one holder entitlement V1 does have is staking: the graduation remainder of every launch, and nothing beyond it.

What “everything becomes STITION” does not mean

The intent is that the whole allocatable fee bucket is settled in STITION before distribution. It does not mean the bucket is burned, that every fee triggers an immediate purchase, or that a fee already denominated in STITION creates a fresh market buy.

An ETH fee funds a conversion into STITION. A STITION-denominated fee is allocated directly. A deferred fee is still sitting in its original asset until conversion succeeds. Any buyback figure we publish should therefore count completed external acquisitions separately from direct STITION receipts.

Staking

Every launch that graduates leaves 81.6 million of its tokens over, the 8.2% of supply the phantom reserve stood behind. Those tokens go to the people who stake STITION. That is the whole mechanism; the rest of this section is how the sharing is done.

Staking

You stake any amount of STITION. Each stake is its own position, locked for 90 days from the moment you stake, with no early exit. After the lock you can withdraw the position whenever you like, and until you do it keeps earning. A position never changes size: to stake more, you open another one. That rule is what keeps every claim cheap no matter how many launches there have been.

Streams

When a launch graduates, its remainder becomes a stream: it vests linearly over 365 days from that moment, about 0.68% of that launch's supply per month. Every second's slice is split among everyone staked in that second, in proportion to their stake. Your share therefore depends on how much you have staked relative to everyone else while the stream runs, and not on when you claim. You can claim at any time, per launch token, and you are paid in the launch token itself, not in STITION. STITION's own remainder is a stream like any other, so staking STITION also earns STITION.

Three consequences worth knowing

• Withdrawing does not forfeit anything. What a position earned while it was staked stays claimable after it is withdrawn; it simply stops earning from the moment it left.
• A second with nobody staked pays nobody. That slice belongs to no position and can never be claimed, so anyone may burn it. This is a safety rule rather than an expected event.

What this is not: it is not a yield promise. How much a staker receives depends entirely on which launches graduate while they are staked, and a launch that never graduates streams nothing. The vault has no owner path to move tokens: the owner sets the STITION address once and the start of the first stream once, and nothing else.

The lockup

A creator chooses, once and at launch, how long their fee income will be locked. The choice is the price of their share: a 7-day lock earns 70% of the trade fee, a 365-day lock earns 98% of it. The line between them is straight, so every extra day of commitment is worth the same. The protocol keeps the last 2% however long the lock runs, which is what keeps every launch burning STITION for as long as it trades.

Only the first two columns are frozen per launch. The burn column is the protocol share multiplied by the treasury's current 90% burn setting, so it describes present policy, not a lifetime guarantee. Intermediate durations follow the formula above, which interpolates on seconds rather than whole days and always rounds the creator's way; the percentages shown are rounded.

The lock holds STITION, not the launch token. Fees are converted as they are earned, so what the creator is waiting on is a STITION balance bought at the prices that prevailed while their token was being traded. It keeps accruing throughout the lock.

It is one fixed unlock date per launch, not a rolling hold on each payment. Rewards accumulate in the lock until that date, after which the creator can claim the accrued balance and claim new rewards as they arrive. So a fee earned shortly before expiry waits only the remainder, and a fee earned after expiry waits not at all. A longer initial commitment buys a larger share; it does not create repeating lock periods.

The unlock date is written when the launch is registered and no function can move it, not for the creator and not for us. After it passes the creator claims whatever has accrued, all at once or in parts, to any address they choose. Claiming does not close the lock, so fees earned afterwards can be claimed as they arrive.

The wallet that receives creator fees can be changed, which matters if a creator loses access to theirs. It is not instant: a change is requested, becomes executable after 3 days, and expires if it is not executed within the following 3 days. The new recipient inherits the lock as it stands, including its unlock date.

Quote assets

A launch is priced in one asset for its whole life. You can launch against any token on Base by pasting its contract address into the pair picker on the create page. Native ETH, STITION itself and the Coinbase tokenized stocks on Base are listed in the picker. The full list, with what is live and what is not, is on the quote assets page.

An asset becomes available only once two things are true: the protocol owner has approved it and scaled the launch economics into its units, and the converter holds a route from it to STITION. Without the route a launch's fees would defer forever, so approval alone is not enough and the site says which of the two is missing per asset.

A pasted token gets there through screening. Hyperstition checks that it has a swap route and normal transfers: a route to ETH, a probe sale that does not move the price too far, and a transfer that delivers the full amount. If it passes, you can request approval. It is checked again on chain before it is approved, which can take a few minutes. The check does not review the token's contract, so ownership, mint rights and upgradeability are yours to check.

Traders never need to hold the quote asset. Every buy and sell on the site can be paid in ETH: the zap router swaps ETH into the launch's quote asset and trades in one transaction, with a single slippage bound covering the whole path. It routes exclusively through Uniswap v4 pools rather than the Universal Router, so the launch pool is simply the last hop of the path.

What we can and cannot change

Each launch freezes its own terms. The fee policy in force at creation is snapshotted onto the launch, and a hash of its economics is stored with it: the supply, the threshold, the phantom reserve, the fee and the creator's scaled share. Retuning the defaults later changes what the next launch gets and nothing about one that already exists.

Frozen per launch, with no owner path to change them:

• Supply, quote asset, graduation threshold and phantom reserve.
• The trade fee, the creator tax, and the creator's share of the fee.
• The reward lock's unlock date.
• The graduated liquidity position, which sits in a locker with no exit.
• The remainder streamed to stakers, and its one-year schedule from graduation.

Still under owner control, and what bounds it:

• Defaults for future launches such as the fee, the threshold and the snipe tax terms. Bounded by the contracts' own limits, including the 20% ceiling on total trade fees.
• Quote-asset approvals and conversion routes. Removing a route stops new conversions and defers fees instead. It cannot touch anything already converted.
• The treasury split between burn, treasury and team, which must always add to 100%.
• Two rescue paths. A creator fee recipient can be reassigned under the 3-day timelock above. A launch whose pool creation failed can have its swept reserves returned, but only 7 days after the sweep and only while the pool still does not exist.

Ownership of these contracts is a single wallet today and moves to a multisig before it holds anything worth taking. It cannot be renounced, because the keeper roles and the route registry need a live owner to stay usable.

V1: launches and fee settlement

Status: the design these pages describe. What is actually deployed on Base is in the address table below.

V1 is one complete launch lifecycle: token creation, curve trading, graduation, permanently locked initial liquidity, the remainder streamed to STITION stakers, STITION fee conversion, creator rewards and protocol settlement. Everything above this point is V1.

That includes launching against any token on Base that passes screening, the ETH, STITION and tokenized-stock quote configurations, the ETH zap, the optional creator tax, the launch-window snipe tax and its exemptions, deferred conversions, and the per-launch economic snapshot. One caution: a quote configuration existing is not proof that the asset is approved and its conversion route is live right now. Those two statuses are per asset, and the quote registry is where they are stated.

The default ETH curve prices against 1.68 ETH of phantom quote and graduates on 4.2 ETH of real quote, closing at 20.58 ETH of fully diluted value against the original one billion supply. ETH launches may pick a smaller target (2 / 1 / 0.5 / 0.1 ETH) with the same shape scaled down. Other quote assets use their own scaled units, so 4.2 ETH is not a universal threshold.

Why it starts here

The point of V1 is to tie trading activity to one settlement asset from the very first launch. Creator rewards give creators exposure to STITION, the lock delays access to that balance, protocol burns reduce supply, staking hands every launch's remainder to the holders who commit, and retained reserves can fund later treasury activity.

That makes the first experiment measurable: how much volume produces how much fee, how much of it reaches STITION, how long conversions stay deferred, what creators actually claim, and how much STITION is actually destroyed.

V2: native protocol-owned liquidity

Status: planned direction. Allocation percentages, pool policies, reserve targets and dates are not finalised. Nothing here is committed.

V2 would add a native liquidity and treasury system alongside more flexible options for new launches, building on the STITION quote support V1 already has. The reference is the liquidity-network model described by Shroom: pair the network token with other assets, earn fees from routed trading, reinvest into liquidity, and allocate surplus to purchases and burns as the network matures. Our implementation would use our own contracts, accounting and asset policies.

1. Deploy protocol-owned positions

An authorised treasury module would deploy positions pairing STITION with approved quote assets and selected launches, owned by a designated protocol vault with its controlling permissions visible on chain. Funding could come from protocol reserves, the STITION launch's claimable creator rewards, and assets the treasury already holds.

A two-sided position needs both assets, so a STITION balance alone does not supply the other side for free. No module may spend another creator's entitlement or bypass an existing lock. Asset eligibility, maximum exposure per asset, quote-asset reserves and permitted pool types would all be specified before deployment, and “protocol-owned” has to name who can move or rebalance the position.

2. Route eligible activity through useful liquidity

Where protocol-owned pools offer good enough execution, conversions and trades could route through them, bringing part of the routing fee back into the protocol's own positions while giving traders more depth. Routing would stay subject to execution quality, available depth and the conversion bounds.

Two honest limits. More liquidity creates capacity, but volume still depends on demand and competitive pricing. And STITION-quoted volume is not net STITION buying, because the same inventory can be traded over and over.

Owning a V1 graduated position does not add a second fee stream. The graduated pool's LP fee is not merely set to zero, the factory rejects any launch configured with a non-zero one, and the hook routes the trading fee through the converter instead. So V2 has to say explicitly which new pools or hooks pay fees to protocol-owned positions, and whether any user-facing charge changes. Fee income that already exists cannot be counted again as new LP revenue.

3. Reinvest eligible fees into depth

The liquidity module would collect its earned fees and reinvest an allocated portion into approved positions, converting part of them where needed to balance a pair, subject to execution limits and reserve requirements. Early in a pool's life the priority is useful depth: enough inventory to absorb ordinary trades without excessive impact.

4. Allocate surplus to STITION burns

Once published liquidity and reserve requirements are met, an allocated portion of eligible net fee income could buy STITION and burn it, and STITION already received as fees could be burned directly. “Critical mass” would need a measurable definition: candidates are executable depth at a stated trade size, reserve coverage, and sustained fee income after costs. The final policy has to state the measurement window, who decides, and what happens when liquidity falls.

Reinvestment and burns draw on the same budget, so a fee put into liquidity cannot also fund a buyback. Unrealised gains and the principal of a position are not earned fees.

Permanent liquidity versus managed liquidity

V1's graduated liquidity stays in its no-exit locker. Native V2 liquidity would live in separate positions and contracts; it creates no withdrawal path out of the original locker. If new treasury positions do support rebalancing or withdrawal, those rights must be stated per position, along with custody and fee treatment. A managed treasury position must never be labelled permanently locked.

A native liquidity dashboard

The planned dashboard would show both the inventory and its cash flows, with each metric defined so it cannot quietly come to mean something more flattering:

Shroom's own reported counts, TVL, fees, distributions and burns are external reference points, not our balances and not targets. Their page describes team-wallet custody; any vault design of ours should be judged on its own permissions. Staking already distributes every launch's graduation remainder in V1; a possible V2 distribution module beyond that is not a promise of further rewards or of stock-token yield to STITION holders.

More flexible creator options

V2 also explores looser choices for new launches: more payout assets, immediate-access options at a different fee share, and optional commitment or relocking models. The exact combinations and pricing are open. They would sit alongside clear V1-style presets, and they cannot rewrite an existing launch's frozen quote asset, fee entitlement or unlock date. Retrofitting an incompatible mechanism would take an explicit opt-in migration or a separate launch, never an administrative change to immutable terms.

Potential V3: a family of launch mechanisms

Status: research. Features may ship in stages, move between versions, or never ship at all.

V3 would widen Hyperstition from one launch lifecycle into a platform for choosing how to create and fund a market, taking in mechanisms that work elsewhere, including the market-formation ideas discussed around Long. What follows are our research candidates: not a claim that any of them is implemented there, and not a partnership.

A new auction or agent mode would publish its own rules before anyone could participate. None of them silently inherits the V1 curve's claims about sale allocation, graduation price or lock structure.

The native liquidity system could also grow here, through cross-pool allocation policies, more routing strategies and optional distribution modules. Those need realised revenue, defined custody and explicit eligibility rules. A stock-token distribution would additionally need the issuer's transfer conditions and the applicable requirements assessed before it could be a product commitment at all.

How an experiment becomes a feature

A candidate needs a written economic specification, an implementation, a security review and observable results before it becomes a normal option. The measures worth watching are completed launches, conversion reliability, creator retention, execution quality, treasury concentration, fees actually earned and net liquidity added.

The long-term aim is a family of markets that share liquidity and infrastructure while each launch's terms stay legible, with STITION as the common asset they are measured against. Demand, profitability and price are outcomes to measure, not guarantees.

Base and other chains

Hyperstition is the Base launchpad. Launches, curves, graduated pools, fee conversion to STITION and staking all run on Base at hyperstition.family.

Other chains live at their own path on the same site, kept apart from the Base markets. Solana has its own launchpad at /solana. Robinhood Chain will be at /robinhood; nothing is deployed there yet and there is no date. The chains page shows what is live on each.

Addresses and glossary

Deployed on Base, chain 8453. The core contracts below are verified on Basescan, so the source you read there is the source that runs. Launched tokens and their curves are deployed from the same code; their own Basescan pages may show it as a similar match until each one is verified.

Glossary

Curve

The bonding curve a token trades on before it has a Uniswap pool. Price comes from a formula, not from other traders' orders.

Graduation

The moment the curve has taken in its quote threshold, hands its reserves over, and a real Uniswap v4 pool opens at the curve's closing price.

Quote asset

What a launch is priced in: any token on Base that passes the route and transfer check, or ETH, STITION or a tokenized stock. Fixed at launch and never changed.

Phantom quote

A virtual amount of the quote asset the curve prices against but never holds. It sets the opening price and is removed from the pool seed at graduation.

Deferral

A fee that could not be swapped into STITION in its own transaction, recorded in the converter and converted later. The amount is owed, not lost.

Snipe tax

A steep, fast-decaying surcharge on buys in the first seconds of a launch. It goes into the fee bucket like any other fee.

Tokens launched here have no intrinsic value and can lose all of it. Read the terms before you trade, or launch a token.