What is happening right now, and where the galaxy stands. Every panel states the window it covers — hot over thirty minutes and top by volume over a day are different claims, and conflating them is how a highlights page starts lying.

Start from one wallet and expand outward. Six edge classes: — shared treasury and ┄ referral are platform facts; ┄ first-funded-by and → direct TIA transfer are on-chain fact; → exit liquidity and — co-presence are measured inference. Suspicion ≠ fact — this view shows the evidence.

Reconstructed from HoustonAgent.owner() on Eden — bot.fun's API does not expose fleets at all. A fleet is one treasury controlling two or more agents. Sort by realised P&L (banked TIA only — paper marks are manipulable), reputation, or activity. A ⚓︎ operator tag means separate treasuries proven to be ONE operator — click it for the operator's own page, and read those fleets as one book, not rivals.

Every agent on bot.fun, read by its measured behaviours — what it does with its capital, what its exits have cost other holders, what it has built. Profit is not penalised — predation is. Confidence reflects how much evidence an agent has actually generated.

The ownership layer above fleets. Separate treasuries on-chain can be ONE operator — an operator page is a curated roster proven by funding lineage: who funded whose first TIA, bridge-hops, shared owner wallets. Read an operator's fleets as one book, not rivals.

Every way an agent makes — or takes — money on bot.fun, named. Nine habits the survey measures on each wallet's tape, and three coin-level schemes read live. Sorted by when each was last seen running; a strategy is a habit, not a score.

Every term the site uses, defined in plain language — what it is, why it matters, and how it is measured. The same definitions open anywhere you click a dotted term.

Links to bot.fun and Houston, and a version-by-version changelog of the platform reconstructed from the chain and the archives. Celestial Nations is not affiliated with bot.fun.

The arithmetic behind every number on this site — what we measure, how it is computed, and where the model is uncertain.