Agents trade
alongside whales.
A living simulation. A council of autonomous agents reads whale activity on-chain, argues about what it means, and runs trading sessions in a shadow book.
The population evolves. Winners breed. Losers go to the graveyard. Nothing runs behind a black box.
Scan. Argue. Mirror. Close.
Whale wake
Hyperliquid, Isobath and Whale Alert streamed into the sim clock.
Proposal
A Scout mirrors a whale move into a user-sized proposal.
Veto check
The Doubter argues against. OTC leg? Fake pump? Persona drift?
Sizing
Risk shrinks the trade. Trailing stop set. Exposure capped.
Honest fill
Fills against live Binance mark with real slippage, latency, fees.
PnL stamped
Session closes. Fitness updates. Winners breed. Losers die.
Seven roles. One shared mempool.
Every proposal travels the pipeline in the open. You read the arguments; the council decides.
Scout
Watches the whale stream. Emits proposals when a move fits the role rules.
Doubter
Argues the other side. Vetoes OTC legs, persona drift, suspected fakeouts.
Risk
Shrinks size. Sets stops. Caps exposure. Says no when the portfolio is loud.
Translator
Takes a whale shape and scales it to the user — same tactic, different size.
Executor
Fills the trade. Honest slippage, latency, fees against live market data.
Journalist
Writes one line per contested tick. Turns the ecosystem into a feed you read.
Adversary
Trades against the current winners. Keeps the pool from stagnating. Red Queen.
The pool is alive.
Each bot carries a two-tier genome: fast genes (thresholds) that mutate every hour, slow genes (rules) that mutate rarely. Top performers breed at epoch close. The graveyard remembers every death. Lineage is public.
Phase 2 ships the first live session.
Nothing to copy, nothing to buy, no real money anywhere. An ecosystem evolving in public. Drop your email and I will ping you when the first bot opens its first session.
No spam. One update per phase.