Everything REALM does, in the order you need it. No jargon that is not explained on the same line.
REALM is a token with a fixed supply. You buy it on the open market, spend it to raise a Castle, and from then on the Treasury pays you a percentage of what you raised, every day, in REALM.
The Treasury is two vaults, filled by people raising Castles and by the tax on collecting. It is not filled by printing: there is no mint function in the token. That means a vault can run out, and it means it can never be inflated away. The number that says how close each one is to running out is its runway, and both are the first thing on every page of this site.
There are two doors. Swear your Castle to the Crown and it pays the Crown’s own rate from the Crown vault: 3% a day for the first fourteen days, then 2%. Swear it to a House — one of ten listed stocks — and its rate is reset at every 4pm close from that stock’s 30-day move — a gain counted five times, a loss twice, never below half what the Crown pays and never above twice it — paid from the Houses vault. You choose once, when you raise.
On the open market, at the price the market gives you. There is no presale, no allocation, and nothing set aside for the builders. Supply is fixed at deployment and cannot grow.
Spend REALM and swear the Castle to the Crown or to a House. Ninety parts in a hundred go to that door's vault and ten are burned. None go to the builders at the raise; article 5 says what the builders may draw from a vault afterwards. Your Castle is recorded at the full amount.
The REALM leaves you permanently. This is not a deposit you withdraw.
From the twenty-fifth hour your Castle earns its rate each day: the Crown’s rate in the Crown (3% a day for the first fourteen days, then 2%), whatever the Court set in a House. Collect it and pay a 25% tax, or fortify it — leave it in — and the Castle grows while you climb the Orders.
You can never collect more than one day's accrual at a time. Nobody can save up a month of coin and dump it in one block.
For the first 24 hours your Castle is scaffolding: it earns nothing and you can abandon it for 90% back. The 10% tithe is burned at the moment you raise and does not come back — it goes to nobody, including us.
After that the way out is to sell the Castle. It is an NFT; whoever holds it receives the stream. The Treasury does not buy Castles back.
The Crown is the default. Its rate is its own — 3% a day for the first fourteen days, then 2% — and no stock moves it. Most people should start here. The Court and its ten Houses (NVDA, AAPL, MSFT, TSLA, AMZN, GOOGL, META, AMD, PLTR, COIN) open with the Realm, each with forty days of price history already proved on-chain, so a House pays its own rate from its first close rather than the Crown’s while a window fills.
A House is a bet on one listed stock, carried by your Castle. Every trading day at the 4pm close, the Court sits and resets each House's rate by one formula:
rate = the Crown’s rate × (1 + leverage × the stock’s 30-day move) — the leverage 5× on a gain, 2× on a loss, the result held between half and twice whatever the Crown pays
| The stock over 30 days | House rate | Against the Crown |
|---|---|---|
| +18.4% (House NVDA today) | 5.76% | nearly twice as much |
| +1.2% (House AMZN today) | 3.18% | a little more |
| −15.3% (House COIN today) | 2.08% | a third less |
| +20% or more | 6.00% | the cap: twice the Crown |
| −25% or worse | 1.50% | the floor: half the Crown |
Between closes the rate is frozen, so weekends and nights pay whatever Friday's close set. The move is read from the stock's on-chain price feed; nobody on our side types a number, and nobody can. The full standings, updated at every close, are on the Court page.
Where the close comes from. Each House's stock has a Chainlink price feed on this chain that publishes on every half-percent move. The Court takes the last print at or before the close and stores it; by the feed's own rule that print is within half a percent of the exchange close. Nobody on our side supplies a number, and the feed's history is public, so anyone can recompute every rate the Court has ever set.
Why a gain counts more than a loss, and why the limits are multiples. Under the symmetric formula this replaced, a House paid about the same as the Crown on average and beat it barely half the time — there was no reason to carry a stock’s swings for it. The limits used to be fixed percentages, 0.5% and 5%, and those quietly changed meaning every time the Crown’s own rate moved; as multiples of the Crown they keep it. Both dials are bounded in the contract and both take 72 hours’ notice, and the 5× is already at the ceiling the contract allows, so it can only ever be lowered.
The forty days behind a House. A House needs thirty days of closes before its move means anything. Rather than pay the Crown’s rate for a month, each House opens with forty days already proved: one feed round per day, each shown to be the last print at or before that day’s close — and a day with no print at all, a weekend or a holiday, proved empty the same way, so no day can be skipped or chosen. Anyone can do it and anyone can check it; until a House is proved in full it pays the Crown’s rate and its older prices are ignored.
No switching. A Castle belongs to the Crown or to one House for its whole life. If you want both, raise two Castles.
The Crown vault pays Crown Castles and is filled by Crown raises. The Houses vault pays every House Castle and is filled by House raises. Neither can pay the other's Castles. Each has its own runway and its own governor, so a run of hot Houses cannot drain the Crown, and a quiet Crown cannot prop up the Houses.
A vault pays out of what it holds. Every day it loses the coin people collect, tax included — the 25% tax is burned, not paid to anyone; it gains ninety parts in a hundred of every new raise. Coin that is fortified never leaves — it grows the Castle, and therefore the future liability, but it does not drain the vault today.
What a vault can pay in a day. A vault releases at most 5% of what it holds in any 24 hours. That is the one speed limit nobody can lift, us included: the vault has no owner and no upgrade. If more than that is asked for in a day, the first collectors are paid and the rest carries over to the next day — nothing owed is lost. The Treasury page shows, for each vault, what it can pay today beside what its Castles accrue today, because a runway cannot show a queue. The marketing draw has a limit of its own, also written into the vault, and uses none of this 5%.
A runway is the vault divided by what it pays out on a normal day. It answers one question: if nobody raised another Castle from this moment, how many more days could everyone in this vault be paid?
Two things follow, and we would rather say them here than have you work them out later:
The governor. No rate is fixed. Each vault pays in full while its runway is above 30 days. Below 30 it halves every rate it pays, below 15 it quarters them, and it restores them as raises refill the vault. A House at 4% in a vault under 30 days of runway pays 2%. This is automatic and written into the contract. Universe cut its rate twice in fourteen days by announcement, after the fact. Ours is a published rule that everyone can see coming.
Everything the builders can do is bounded by range, by notice period, and by destination — and the boundaries are in the contract, not in this document. The full list, with the live Admin Log, is on the Treasury page.
The short version: we can move the Crown rate inside a published range with notice, set the House leverage each way (1×–5×) and its floor and cap (a quarter to two and a half times the Crown) inside those ranges with notice, add a House, set the tax inside a published range with notice, draw up to 1% of a vault for marketing at most once every 6 days, announced 72 hours ahead, pause for at most 48 hours, and refund a Castle to its own owner. We cannot mint, cannot take anything from a raise, from the collect tax or from a vault, cannot set a House's rate by hand, cannot change what a Castle is owed, cannot freeze a wallet, and cannot block the 24-hour abandon.
REALM is a deliberate rebuild of a design that failed publicly, and it is only fair to say so on our own site.
Universe launched on Avalanche on 25 January 2022. You burned its token to mint an NFT "Planet", and the Planet paid you 4% a day, forever. It was better built than its competitors: a real product, more than 45,000 Planets in two weeks, a treasury genuinely deployed into Axial and BenQi that anyone could check on Zapper, a candid changelog, and a real third-party audit delivered ahead of schedule.
It cut the rate from 4% to 3% eleven days after launch, and again to a tiered ~2.1% three days after that, both times because it was growing faster than it could pay. By the end of April the price was down about 99.5%.
Then, on 24 April 2022, the team upgraded the proxy contracts six times in twelve minutes and nine minutes later minted 16.2 trillion tokens to a fresh wallet. Over the next three and a half hours they minted 11.5 quadrillion in total, drained both liquidity pools in around forty-five declining slices, and left with about $269,776. The final implementation they installed has never been verified and is still the live one today.
The mechanism was not an exploit. The token contract was genuinely immutable and said so — but it hard-coded its unlimited mint permission to an address that was an upgradeable proxy. The audit had flagged exactly this under "centralization" and still returned an overall PASS.
What we changed because of it. No mint function exists in REALM at all, so the worst case is an empty vault and never an infinite supply. Neither vault has a proxy, an owner or a withdraw function. The daily-limit logic lives in the Castle contract, not inside the token's transfer hook — Universe's equivalent was the first thing they replaced on the night. And the runway, the one number Universe never published, is on every page here.
Our full transaction-level reconstruction, including the exact upgrade timeline and the wallet that received the mints, is in the research this project was built from.
A vault can run out. That is the central risk and it is not hidden anywhere. If raises slow, its runway falls, the governor cuts its rates, and if it keeps falling the vault empties. Your Castle would then be worth whatever someone else will pay for it.
Later is worse than earlier. Anyone who raises early recovers their money sooner. Anyone who raises late may not recover it at all. The runway on this page is the honest way to judge which one you would be.
The token's price can fall. Your Castle pays a fixed number of REALM a day. If REALM halves, so does the dollar value of every payment you have not yet taken.
Fortifying is a bet on the future. Coin you fortify is coin you did not take. It grows your Castle, and it is only worth something if its vault is still paying later.
A House is a leveraged bet on one stock. A stock down 15% over a month costs its House about a third of what the Crown pays; down 25% puts it on the floor, half the Crown, for as long as the slide lasts. A gain counts for more than a loss, but the loss is real. You cannot switch out. If you do not have a view on the stock, take the Crown.
The builders can draw from the vaults. Up to 1% of a vault, at most once every 6 days, which is no more than 5% a month, announced on-chain 72 hours before and paid to one named wallet. That limit is in the vault itself, which cannot be changed, though we may draw less. It binds the draw, not our code: the vault cannot tell who a payment is for, so an upgraded version of our code could pay anyone out of the 5% a day holders are paid from, and what limits that is the 5% itself and the seven public days every upgrade must wait. It uses none of the 5% a day paid to holders. It is for marketing, and nothing in the contract makes us spend it on that. Every REALM drawn is one the vault cannot pay to a Castle, and drawn REALM that is sold is sold into the same market you sell into. The builders are also paid the trading fee on the market, in USDG, which does not touch a vault.
The Houses vault is shared. Every House is paid from one vault. A month where several Houses run hot drains it faster than the Crown, and the governor halves every House when its runway drops under 30 days, the cold ones included.
The rate is not a promise. A vault pays in full while its runway is over 30 days, halves under 30, quarters under 15. If you are counting on today's rate continuing, you are counting on other people raising Castles.
Your Castle is only as liquid as the next buyer. After the first 24 hours the Treasury does not buy Castles back. The exit is the secondary market: a Castle is a standard NFT any marketplace on the chain can list, and its deed shows its House, Order, value and today’s rate, read live. Selling it moves nothing out of the vault. A Castle sells for what someone will pay for a stream from a vault they can see the runway of.
| Version | Date | What changed |
|---|---|---|
| v1.0 | — | First publication. Crown 2.00%/day, steady · House rate = the Crown’s rate × (1 + leverage × 30-day move), leverage 5× on a gain and 2× on a loss, held between half and twice the Crown, reset at each 4pm close; each House opens with forty days of closes proved on-chain · Crown opens at 3.00%/day for 14 days, then 2.5% for a day, then 2.00% · every raise 90 to its vault and 10 burned, nothing to the builders at the raise · a marketing draw of at most 1% of a vault per 6 days, 72 hours’ notice · two vaults, each with its own runway · collect tax 25%, burned · collect cap one day's accrual · the Order bonus climbs at most one step a day · every Castle carries a deed marketplaces can read · governor halves under 30 days, quarters under 15 · no switching · 24-hour abandon at 90%. |
Every future change to a published number appears here as a numbered version with the old and new value side by side, and goes through the notice period on the Treasury page first.
Every figure on this site is example data for a design in progress. Live numbers come from the contracts once they are deployed.