Modes and strategy

Buying is identical in every mode: your bot buys what our alpha bot buys, at your stake. The mode decides how it sells — and selling is where memecoin outcomes are actually made and lost.

All three modes run on the same exit machinery our own bot uses — the same ladder, the same price feed, the same code. A mode is a preset over that machinery, not a separate strategy that can quietly drift from it.

Normal

Our exact current approach. The protective stop, both take-profit rungs, the time limits, the trailing exit, the panic and dead-volume exits, and the 10% moon bag — all as our own bot runs them, including every tuning we ship in the future. If we improve our exits, your normal bot improves the same day, automatically.

This is the default for a reason: it is the only mode where somebody is actively measuring the strategy against outcomes and adjusting it.

Read exits and the ladder for what each exit actually does — everything there applies verbatim to a normal-mode bot.

Aggressive

Take-profit steps only. TP1 and TP2 fire; nothing else ever sells. No stop-loss, no time limits, no trail, no dead-volume exit, no moon-bag clamp. Winners bank on the way up; whatever remains after TP2 rides indefinitely — to a 100× or to zero.

Be clear-eyed about what this is: most fresh memecoins go to zero, and an aggressive bot has no downside protection at all. A token that never reaches TP1 loses the entire stake, every time. This mode exists for people who want exactly that trade-off and say so.

Custom

Your own strategy, defined per cohort. Every token our bot buys is stamped with how fast it graduated, and your custom strategy sets a separate plan for each band:

Cohort	Graduated in
instant	under a minute
fast	under an hour
mid	under ~8 hours
aged	longer than that

The bands behave differently in practice — an instant graduation is usually a funded team filling its own curve; an aged one is a slow burner — which is why the plan is per-cohort rather than global.

For each cohort you set:

Entry plan

• buys / skipped — whether the bot enters this cohort at all. A skipped cohort's tokens are simply never bought.
• stake × (0.25–2) — scales your base stake for this cohort. A cohort you half-trust can ride at half size. The result still respects the platform stake ceiling.

Take-profit ladder

• TP1 at × (1.05–10) — the first rung's trigger, as a multiple of your entry.
• TP1 sell % (10–100) — how much of the position banks at TP1.
• TP2 step × (1.1–10) — the second rung, as a step off where TP1 actually filled — not off the configured trigger. A rung set at 1.25× can fill at 1.38× when the price gaps; TP2 is measured from the fill, because that is the price you actually got.
• TP2 sell % (10–100) — how much of the remainder banks at TP2.
• trail — after the rungs, a trailing exit protects what is left from round-tripping a peak.

Protective exits

• stop-loss — the hard stop under your entry.
• time limits — the cuts that close a position going nowhere: the pre-TP cut, the post-TP1 cut, and the banded time stop.
• panic & dead-volume — sells into a rug-shaped tape, and closes tokens whose volume flatlined.

Each toggle disarms exactly one family and leaves the rest alone. Turning everything off and keeping only the TP ladder rebuilds aggressive mode by hand — with the difference that here you chose it per cohort.

Moon bag

Whether the bot keeps the protected 10% remainder our own strategy keeps — the slice that stays for the tail — or sells positions to zero.

Two properties worth knowing

• Every number is clamped server-side. A TP1 of 0.5× (a take-profit under water) becomes 1.05×; a stake × of 99 becomes 2. The form cannot produce a strategy the ladder would refuse.
• Edits apply to open positions. Changing a cohort's plan re-steers positions already open in that cohort from the next tick — the same live behavior as our own operator panel. Anything you leave untouched is our strategy's own number.

Launchpads

Each bot mirrors only the launchpads you select — pump.fun, LetsBonk (Raydium LaunchLab), StonkFun. The filter resolves through the token's actual on-chain launch venue, and an unknown venue matches no bot, never silently counts as pump.fun.

ansem.io appears in the list as a curated channel rather than a launchpad of its own: its projects launch on pump.fun's program, and we identify them by the launch wallet's funding. Picking pump.fun includes them; picking ansem.io alone mirrors only ansem-curated launches.

One honest caveat: our alpha bot currently trades pump.fun graduations. Selecting another launchpad prepares your bot for the day our book trades there — it does not conjure signals we are not producing. The default (pump only) is the setting that does something today.

What is deliberately NOT mirrored

• DCA / averaging down. When our bot averages into a falling position, it is a decision about our basis and our stop. Your bot never doubles its stake on a falling token — its wallet holds a fixed budget you chose.
• Late-signal reprieves. Our bot occasionally holds a time-limit cut when live trade flow, holder growth, or a paid listing argues for patience. These signals are wired to our own positions; v1 user bots take the plain cut instead. Conservative in the literal sense: it sells sooner, not later.