Where winning a battle literally moves market cap from the loser's token to yours.
Every bonding-curve product ends up as the same loop, and that loop only holds active traders. The team wanted a reason to open the app on a flat day, and concluded that pure speculation is a retention dead end no amount of badges would fix.
Each launch is tied one-to-one to an AI Monster running on an AI Town-derived agent simulation, with its own behaviour patterns rather than static stats. Its bonding-curve Shard token is bound programmatically to that Monster's identity and progression, so the asset cannot be separated from the creature.
In the Training Dojo, players run waves against AI opponents to earn experience, unlock moves and evolve forms. Training does not just raise numbers — it refines the AI's decision-making for future battles, so investment of time shows up as smarter play.
Battles resolve entirely by AI, and the winner takes a defined portion of the loser's market cap. Price action becomes a legible, consequential event instead of a chart. A weekly points leaderboard pays USDC to the top fifty, scoring battles, training consistency, trading, referrals and community activity so the incentive is not volume alone.
“They made the number going up and the monster winning the same moment.”
Built on Abstract rather than Solana, the prototype is our clearest example of binding a curve-priced asset to a game with its own reason to exist.
