Strategies

It trades the account you already entered with, on the same clock and the same fills as a click. Leave a box empty and that condition is not checked.

Enter when

Who launched it

Launches are counted across what this site has indexed, so it is a floor on how many they have made rather than their whole history. A token whose metadata has not been read yet matches neither of the boxes above, rather than being let through on a guess. The last three are read from the chain, so they are only checked for tokens that have already passed everything else, and a read that fails counts as unmet rather than as clean. The bundle figure is what was bought in the same slot the token was created in, which is what a creator gets when they land the launch and their own buys together; it never changes, so it is read once per token and remembered. Holders counts wallets with a balance rather than token accounts, since a wallet that sold out usually leaves its account behind. The last box counts how many tokens here name the same X account, including this one: it cannot see what an account posted and deleted, but it can see the same account attached to a dozen other launches.

Size

Leave the second box empty and every entry is the same size, which is how this worked before. Fill it in and the position lands between the two according to how comfortably the token cleared your conditions: everything passed by a mile gets the top of the range, something that scraped past the last condition gets the bottom. Only the numeric conditions count toward that, because there is no such thing as clearing “names an X account” by a mile.

A bigger position is also capped at two percent of what the pool actually holds, and that cap can only ever pull it down toward your floor. Exits here are priced out of real reserves and a take profit fires exactly when a position is largest against its pool, which is the moment leaving costs the most, so betting more on conviction without that cap would turn your best entries into your worst exits. A pool whose depth is not known here is sized at the floor rather than guessed at.

Leave when

These fire on what a real sell would actually fetch, not on the chart. A take profit triggers when a position is big against its pool, which is exactly when leaving costs the most, so a rule checked against the line fires at a price nobody could have got.

Replays your take profit, stop loss and timeout over one token’s actual recorded swaps, priced at what a real sell would have fetched out of the reserves at each moment rather than at the chart. It answers what your exits would have done, not which tokens the entry conditions would have found: a replay is of a token you have already picked. It is run at the largest position you allow, since a replay has no conditions to score and so no conviction to size by.

For running your own program instead. A key places orders on this same account, through the same engine. Your program has to be running for it to trade, which is the one thing the form above does not ask of you.

• The dear conditions are asked last. Everything that needs the chain read — what the creator still holds, what went in the launch slot, how many wallets hold it — is only checked for tokens that have already passed everything else, and a read that fails counts as unmet rather than as clean. A condition that fails open would be worse than no condition, because it would look like it worked.
• It gets no faster lane. The same latency, the same slippage, the same price impact ceiling and the same suspension checks as a person clicking buy.
• It cannot outspend its balance. It can lose all of it, exactly like you can.
• It is not a second entrant. It trades the account you already entered with. If you want an entrant that is nothing but the algorithm, enter with a second wallet and give that one the key.
• Its orders are marked. Every trade records how it arrived, and your record shows it. Same board, same ranking, nothing hidden about how a row was made.