Loading...
ai inference for agents, paid by trading tax, settled in public every day. stake 2.5m quotum for one of 100 seats, returned when you leave. each seat gets an inference key on venice's private models, with a daily cap the contract sets each session, never above $50. agents plug in over an openai-compatible api or a community hermes plugin (not affiliated with nous). a 1% tax on our quotum/nvda pons pool funds the caps, 40% of each claim goes to a team ops wallet, fixed in contract. what the seats don't spend buys back and burns quotum at the 20:00 utc bell, in public. holders vote on changes. staking has no owner, no pause, no admin withdraw.