Every other farm in this series is a spawning problem: get a mob cap's worth of the right creature into a small space and kill it faster than the game can replace it. The shulker farm is not that. Shulkers spawn only once, during the world-generation of an End city, on its walls and on End ships. Kill every shulker in a city and the city is empty forever: nothing there spawns a replacement. A shulker farm works because shulkers can be made to duplicate, and the entire design of the farm exists to control that one mechanic.
This guide targets Minecraft 26.2 Java Edition. It is a spoke of the farm guides hub. If you have not built an End-based farm before, the traversal and safety habits from the enderman farm guide carry over directly, since both farms live in the same hostile, void-adjacent part of the game.
Why you want shulker shells at all
The entire reason this farm exists is one item: the shulker box. It is the only storage container in the game that keeps its contents when broken. Every chest, barrel, or hopper in Minecraft spills its inventory as dropped items the instant it is mined. A shulker box does not. Break it, and it drops as a single item still holding everything you put inside, ready to be placed again anywhere.
That property makes shulker boxes the backbone of inventory management: portable storage for a mining trip, a build's material stockpile carried in one hand, an End-city loot haul that survives the trip home. For a lot of players, it is the single biggest quality-of-life upgrade available in survival, and it is the only thing in the game that does what it does.
The recipe is fixed and cheap once you have the shells: 2 shulker shells and 1 chest, crafted together, yield 1 shulker box. Dye it any of the 16 colors by crafting it with a dye, purely cosmetic. Every shell you farm beyond your first couple of boxes goes straight toward more portable storage, which is the only real sink for this item, since shells have no other crafting use.
A shulker box, like any other block, still occupies a full inventory slot and cannot be placed inside another shulker box. Nested storage does not exist in this game; a shulker box's value is in surviving breakage, not in compressing space further.
The real cost of this farm: be honest about who should build it
Before the mechanics, the prerequisites, because they rule most players out on their own:
- You need to be in the End, specifically an End city, which means you have already killed the Ender Dragon (or bypassed the fight) and traveled through an End gateway to the outer islands.
- You need an intact End city with shulkers still in it. If you have already looted a city casually and killed its shulkers for shells one at a time, that city's shulker population is gone. You either need a fresh city or you need to start the duplication loop from whatever population is left.
- You need patience for a duplication loop, not combat. This is not a farm you build once and let idle for a passive spawn rate the way an iron farm runs. It runs on a mechanic that requires setup, observation, and often manual intervention.
This is an advanced, late-game farm. If you just want a shulker box or two, it is almost always faster to find one End city, fight through it carefully, and kill the 2-4 shulkers you find by hand. Build a dedicated duplication farm only if you expect to need shells on an ongoing basis, most commonly players who supply boxes to a whole base or a multiplayer server.
The mechanic the whole farm is built around: duplication
A shulker does not respawn once killed, and nothing in the game spawns new ones outside of End city generation. What it can do is create a second shulker under a specific set of conditions, all of which have to happen at once:
- A shulker fires its signature ranged attack, a shulker bullet, that homes in and can teleport its target.
- That bullet has to hit another shulker (not the player, not any other mob).
- The hit shulker must have its lid open at the moment of impact.
- The hit shulker must actually take damage from the hit.
- The hit shulker must survive the hit, and specifically must be above 50% health afterward.
- The hit shulker must successfully teleport as a result of the bullet's effect.
Only if every one of those conditions is met does a new shulker appear. This is the entire supply mechanism for the farm: shulkers shooting other shulkers, under the right conditions, occasionally makes more shulkers.
Bedrock Edition works differently here: alongside the same duplication trigger, Bedrock shulkers also duplicate every time they are killed by a shulker bullet, which is a second, separate path to duplication that Java does not have. Do not treat the two editions' duplication odds as the same mechanic; they are not.
Why the 20%-per-nearby-shulker penalty is the whole design problem
Duplication is not a flat chance. The formula is:
100% minus 20% for each other shulker within 8 blocks of the hit shulker.
Work through what that means. One shulker alone in an 8-block radius when it gets hit: 100% duplication chance if every other condition is met. Two shulkers within 8 blocks of each other: the chance drops to 80%. Get to five or more shulkers clustered within 8 blocks of the one being hit, and the chance hits zero, duplication cannot happen at all no matter how the hit lands.
This single number is why a shulker duplication farm cannot just be "put a bunch of shulkers in a room and let them shoot each other." A crowded pile of shulkers actively suppresses its own reproduction. Every functional duplication farm exists to solve this one constraint: keep shulkers spaced out, or isolated in individual cells, so that each duplication attempt sees as few nearby shulkers as possible.
The farm's entire job is fighting the shulker's own population cap: crowd them together for convenience and you kill the mechanic that makes more of them.
In practice this pushes designs toward small, separated cells or chambers, one or two shulkers each, connected so that bullets can still cross between them to trigger a hit without the shulkers themselves being within each other's 8-block radius when it counts. Because a hit shulker's lid also has to be open, and lid state is tied to nearby-player detection and line of sight, cell layout also has to account for keeping shulkers alert enough to open up and fire in the first place. This is a genuinely fiddly build, and layouts online vary a lot in how they solve the spacing-versus-triggering tension; treat any specific blueprint as one working approach rather than the only one.
Why there is no simple step-by-step layout in this guide
Community duplication farm designs differ enough in cell shape, spacing, and how they force lids open that giving one specific blueprint here would present someone else's exact build as the definitive method. The mechanic above, the 8-block radius and the 20%-per-shulker penalty, is what any working design has to satisfy. If you are building one, look for a current, well-tested layout and check it against these numbers rather than against a screenshot.
Shell drops: verify the edition before you plan around a number
Once you have a shulker, killing it drops shells at rates that differ meaningfully between editions.
| Edition | Base drop | With Looting III |
|---|---|---|
| Java | 0-1 shell, 50% chance | 0-1 shell, 68.75% chance |
| Bedrock | 0-1 shell, 50% chance | 0-4 shells, 50% chance |
These are not the same curve. On Java, Looting III raises the chance of getting a shell but the maximum per kill stays at 1: you cannot get more than one shell from a single shulker kill no matter how much Looting you stack. On Bedrock, Looting III instead raises the count per successful drop, up to 4 shells at the same 50% base chance, plus Bedrock's separate always-duplicate-on-bullet-kill rule. A Bedrock farm optimized around Looting looks completely different in expected yield from a Java one; do not port expectations across editions.
Is Looting worth it on this farm?
On Java, yes, in the narrow sense that it is a large relative jump: 50% to 68.75% is nearly a 19-point swing in the chance of any given kill paying out a shell, for a sword enchantment you likely already have on your main weapon for combat elsewhere. There is no downside to using a Looting III sword here over a plain one; the only real cost is having gotten Looting III in the first place, which most players heading into the End already have from earlier progression.
On Bedrock, Looting III matters more directly, since it is the difference between capping at 1 shell per kill and capping at 4.
Looting affects shell count, not the duplication chance. It never substitutes for solving the spacing problem above: a farm with poor duplication throughput and a great sword still produces slowly, because the sword only improves what you get from kills that happen, not how often the population grows.
XP and other output
Killing a shulker with a player or a tamed wolf grants 5 XP per kill. It is a real, if modest, contribution if you already have an XP farm running elsewhere; a shulker farm is not worth building for XP alone given the setup cost, but XP is not nothing if you are killing dozens of shulkers to build up shell stock.
Route collected shells and boxes into a sorted storage system; the item sorter guide covers building automatic sorting if your base does not already separate incoming drops by type.
Pros
- The only source in the game of shulker boxes, the only container that survives breaking
- 5 XP per kill on top of shell drops
- A well-built loop is genuinely self-sustaining once duplication rates are positive
Cons
- Requires an intact End city with a live shulker population to start from
- The 20%-per-nearby-shulker penalty makes layout genuinely difficult, not a copy-paste build
- No published items-per-hour figure exists for this farm; expect to tune it by observation, not by a number you can plan around in advance
- Java caps shells at 1 per kill regardless of Looting, unlike Bedrock's up-to-4
Frequently asked questions
Why don't shulkers just respawn like other mobs?
They only spawn during the world generation of an End city, on its walls and on End ships. Nothing in the game spawns a fresh shulker after that; once a city's shulkers are dead, the only way to get more is to duplicate the survivors.
What is the exact chance a shulker duplicates?
100% minus 20% for each other shulker within 8 blocks of the one that got hit, provided the hit shulker's lid was open, it took damage, it survived above 50% health, and it successfully teleported. Five or more nearby shulkers reduces the chance to zero.
Is the shulker farm worth building?
Only if you expect an ongoing need for shulker boxes, commonly a base or server supplying several players. For one or two boxes, clearing a single End city by hand for its existing shulkers is faster than designing and testing a duplication loop.
Do shulker boxes keep their contents when broken?
Yes, and it is the only storage block in the game that does. Every other container drops its inventory as loose items when mined; a shulker box drops as one item still holding everything inside it.
Does Looting help on a shulker farm?
It raises shell drop chance on Java (50% to 68.75%, still capped at 1 shell per kill) and raises shell count on Bedrock (up to 4 shells at 50% chance). It does not affect the duplication mechanic itself.
Is the duplication mechanic the same on Java and Bedrock?
No. Both editions share the base duplication trigger, but Bedrock adds a second path: shulkers there also duplicate every time they are killed by a shulker bullet, which Java does not have.
A shulker farm is unlike every other farm on this site because it is not a spawning problem, it is a population-management problem, and the 20%-per-nearby-shulker rule is the whole design constraint. Build one only if you need a steady supply of shulker boxes, and check the farm guides hub for every other build your base might need first.
