the numbers
| supply | 4,000 eggs, fixed. The contract cannot mint a 4,001st. |
| mint price | 200,000 $HATCH per egg. |
| split | 100,000 burned on every mint. 100,000 into the holder pool. |
| wallet cap | 8 eggs per wallet, lifetime. Selling does not free a slot. |
| reveal | Eggs stay eggs until a hatch event fires. |
why later eggs cost more
The mint price never changes in $HATCH terms. What changes is $HATCH itself. Every egg burns 100,000 tokens forever, so a full sellout burns 400,000,000 of the 1,000,000,000 total supply. Forty percent of the token, gone, purely from minting.
As supply shrinks, the same 200,000 $HATCH represents a larger slice of what exists. Egg number 3,000 costs the same number of tokens as egg number 1 and a meaningfully larger share of the remaining supply. That is the whole pressure mechanism. There is no bonding curve and no price oracle, just arithmetic on a shrinking denominator.
the wallet cap
Eight per wallet keeps any single buyer from owning the reveal. The cap is lifetime and counts mints, not holdings: if you mint 8 and sell 8, you are done minting. You can still buy on secondary without limit. The cap exists to spread mints, not to control the market after them.
traits are locked before you mint
Every egg's contents are generated before the mint opens: species, palette, headgear, item, background. The full list is already generated and hashed. This is the commitment, live today:
sha256(traits.json) =
c5f862160a5c4a25e0fac642c95debef0101d1d2777d79d74f9460be458fda79When an egg hatches, its contents come from that pre-committed list. How to verify after the final hatch is on the docs page.
After the final egg hatches, the complete trait list is published. Anyone can hash it and compare against the day-one commitment. If even one trait had been swapped after launch, the hashes would not match. You do not have to trust that rarity was not edited to favor anyone; you can check.
hatch events
An event fires when Codex makes real news: a new model in the family, a major feature release, an official benchmark result. When one fires, the oldest 400 unhatched eggs crack, in mint order. Ten events fully hatch the collection.
This means early minters hatch first, and nobody, including us, controls the calendar. If OpenAI ships fast, the collection reveals fast. If they go quiet, everyone waits together. The qualifying-event rules are published before launch so a hatch can be argued about in public, not decided in private.
One honest caveat: someone still has to press the button when news qualifies. The hatch trigger is operated by us against the published rules. The trait contents are trustless via the hash; the timing is not.
