Running a fleet without it running you

Decide the roles before you add anything

The single highest-leverage decision in running a fleet is what each account is for, made before you import it rather than worked out later once forty accounts are already running undifferentiated. This matters more than any setting, because a fleet where every account is treated identically is a fleet where you can't reason about risk, priority, or what actually happens if one specific account gets banned. Most working operations sort into three roles, and the sorting itself — not the automation underneath it — is what makes a fleet manageable past the point where you can hold the whole thing in your head at once:

• Farms — disposable, gather constantly, exist to feed something else. Losing one costs a week.
• Banks — hold and forward resources. Fewer, more valuable, lower profile.
• Mains — what the whole operation exists to serve. Most experienced operators never automate these at all; see the ban-risk guide.

Tag on import, not later. Retagging forty accounts after the fact is the kind of chore that never gets done, and every fleet-wide view is worthless until it is.

Proxies: when they earn their cost

Proxies are the most over-prescribed item in this hobby. Reasoning, not measurement:

1. A handful of accounts: probably unnecessary. Normal households have several players on one connection.
2. Dozens on one connection: worth considering, since the pattern stops resembling anything a household produces.
3. Accounts you cannot afford to lose together: the real argument for separation is blast radius, not detection.

Apex takes a SOCKS proxy per account, so you can separate selectively rather than paying for the whole fleet.

How much hardware, roughly

None of yours: Apex runs on our servers, on every plan including the operator plan. For reference, Apex needs 8 GB of memory at minimum — the floor for the console and a small fleet, not a number that scales with account count on its own. What actually grows with the fleet is whatever emulator or client instance runs each account: each running copy holds its own memory footprint, so the real number to plan around is that per-account cost multiplied by how many you intend to run at once, not the 8 GB baseline in isolation.

What actually limits fleet size

Not the software. In order of what people hit first:

• Memory. The usual ceiling, for the reason above. Budget generously and check real usage at your account count before scaling further.
• Upstream bandwidth, if every account is proxied through one connection.
• Your own attention. The one nobody plans for. Past roughly twenty accounts, manual checking stops working — which is the entire reason Apex flags accounts rather than expecting you to go looking.

The operating routine that works

Once tagging and recovery are in place, the daily job collapses to three things:

1. Glance at fleet-wide resource totals. If last night looks like the night before, nothing needs you.
2. Look at anything flagged. Apex gives up after three restart attempts precisely so a genuinely broken account surfaces instead of looping quietly.
3. After a game update, expect breakage and check early. Every bot breaks on update day; what differs is how fast you notice.

If the fleet is a business

Operators renting accounts to other people hit a different wall: every settings change becomes a support request. That is what the client portal is for — licensed operators hand it to their customers, who adjust their own whitelisted settings and Apex applies them without a restart. Pricing is here.