The short answer
A server-grade Minecraft survival island seed has four witch huts inside a 256-block radius, more than one ocean monument, a bastion remnant inside practical reach, and the whole cluster inside roughly 1,000 blocks of spawn. Three seeds on this page passed that sheet on Java 1.18+: -8350249385580253962, 3387664876512721018 and 4188583766222650532.
If you only want one, take -8350249385580253962. It is the only one of the three with a spawn point recorded in the test log, and the tester’s summary line reads: four witch huts, an ocean monument, a bastion remnant and a warm ocean inside 600 blocks of the island spawn. The other two are mapped to structure coordinates, not to a recorded spawn.
Solo seeds and server seeds are judged on different things
Almost every island seed list you have read is written for one player. The criteria are survival criteria: trees, fresh water, a shipwreck, maybe a village close enough to walk to. Those are the right criteria for a single player, because the question a solo seed answers is “can I get through the first night and the first week”.
A shared world asks a different question. Nobody is trying to survive the first night — eight players with a spawn point and a chat window will manage that. What they need is a world where several farms can run at the same time from one base, where the Nether route opens early, and where nobody spends their evening walking 1,000 blocks to reach the thing they were told was “nearby”. That is a different sheet, and it is the one below.
Nothing here was produced by scanning. Every line comes from worlds that were loaded and mapped structure by structure, with the tester’s accept or reject decision recorded alongside the coordinates.
The four checks, in the order they are applied
- Distance budget, checked first. The working cluster has to sit inside roughly 1,000 blocks of spawn. This is a hard fail and it is checked before anything else, because a beautiful structure 3,000 blocks out is not a feature on a shared world — it is a project someone has to build a Nether highway for.
- Four witch huts inside a 256-block radius. Quad witch huts let one farm cover all four spawning areas, which is the most efficient source of gunpowder, redstone and glowstone dust in the game. A seed without them is marked not suitable for servers.
- More than one ocean monument. A single monument is a structure you clear once. Two or four is a structure you farm, for prismarine and gold, forever. This is why one seed with quad witch huts directly at spawn still got turned down: it had no second monument anywhere.
- A bastion inside reach. Early Nether access is treated as a requirement, not a bonus. Without a reachable bastion the gold and blaze rod pipeline does not start, and everything downstream on a shared world waits on it.
Two more things get recorded on every seed even though they sit outside the pass/fail line. Warm ocean access gives coral, sea pickles and tropical fish, and keeps boat lanes usable. And the crossroads — the junction everyone travels through — is marked down when it only has three connections, even when the structures around it are good.
One nuance worth keeping: distance is measured in travel time, not straight-line blocks. A quad monument cluster 3,000 blocks out is not automatically disqualified, because Nether travel compresses the run to roughly 388 blocks. What kills a seed is a whole cluster sitting past the budget with no shortcut to it, which is exactly what happened to -9147459640944998301.
The three seeds that passed
-8350249385580253962
Quad witch huts Quad monument Bastion in reach Warm ocean Spawn recorded
- Spawn -150, 250
- Witch hut -325, 325
- Ocean monument -300, 150
- Pillager outpost 600, 200
- Warm ocean 550, 1100
- Nether fortress -150, 0
- Bastion 100, -100
- Quad monument -3100, -1100
This is the benchmark seed of the three and the only one with a spawn point in the log. The tester’s summary: four witch huts, an ocean monument, a bastion remnant and a warm ocean, all within 600 blocks of the island spawn.
The interesting part is the quad monument at -3100, -1100. On its own that is a disqualifying distance — over 3,000 blocks out. It survived the review because Nether travel compresses the trip to about 388 blocks, which turns a distant cluster into a short hop once the highway exists. Distance gets measured in travel time, not in blocks.
Java 1.18 and above. Nothing here was verified on Bedrock, and the two editions do not share world generation.
3387664876512721018
Quad monument Bastion in reach Warm ocean Nether fortress Spawn not recorded
- Witch hut 200, 150
- Ocean monument -250, 150
- Pillager outpost -225, -325
- Warm ocean 0, -500
- Nether fortress 250, 0
- Bastion -250, 50
- Quad monument -3650, 950
The tightest structure layout of the three: monument 150 blocks off the hut axis, Nether fortress at 250, 0, bastion at -250, 50. The bastion is the closest of any seed on this page, which is the check people most often skip.
The log records a witch-hut coordinate at 200, 150 but no explicit count, and no spawn point. Treat both as unverified until you load the world yourself — the coordinates above are what the tester wrote down, and they are what the acceptance was based on.
Java 1.18 and above. Load it on the edition you play before trusting any of it.
4188583766222650532
Dual monument Mushroom island Two Nether fortresses Bastion in reach Spawn not recorded
- Witch hut 350, 350
- Ocean monument 350, -250
- Pillager outpost -350, 0
- Warm ocean 0, -250
- Mushroom island -300, -500
- Nether fortress (near) 0, -150
- Nether fortress (quad) -100, 200
- Bastion 150, 0
- Dual monument 3850, -4150
The most unusual of the three, because it is the only one carrying a mushroom island, and the only one with two Nether fortresses recorded — one near at 0, -150 and one logged as the quad-side fortress at -100, 200. Two fortresses means the blaze rod route has a fallback if the first one is stripped badly.
Its monument count is the weaker line: the log records a dual monument at 3850, -4150, not a quad. That is still more than one, which is what the check asks for, but it is the thinnest pass on the sheet. The mushroom island at -300, -500 is the compensation — no hostile mobs spawn there, which makes it the safest place on the map for a shared storage base.
Java 1.18 and above. The dual monument is recorded at 3850, -4150, which is far outside the working radius unless a Nether route is built.
The pass sheet, side by side
| Check |
-8350249385580253962 |
3387664876512721018 |
4188583766222650532 |
| Spawn recorded |
-150, 250 |
Not in the log |
Not in the log |
| Witch huts (256 radius) |
Four, within 600 blocks of spawn |
Hut at 200, 150 — count not recorded |
Hut at 350, 350 — count not recorded |
| Ocean monuments |
-300, 150; quad cluster -3100, -1100 |
-250, 150; quad cluster -3650, 950 |
350, -250; dual cluster 3850, -4150 |
| Bastion in reach |
100, -100 |
-250, 50 |
150, 0 |
| Warm ocean |
550, 1100 |
0, -500 |
0, -250 |
| Nether fortress |
-150, 0 |
250, 0 |
0, -150 and -100, 200 |
| Extra features |
Pillager outpost 600, 200 |
Pillager outpost -225, -325 |
Mushroom island -300, -500; pillager -350, 0 |
| Verdict |
Accepted |
Accepted |
Accepted |
Read the “count not recorded” rows honestly. An accepted seed is not a seed where every number is on paper — it is a seed where nothing failed badly enough to stop the review. The log is a tester’s notes, not a scanner’s output, and where a number is missing it is missing in the source too.
What failing looks like
Four seeds were mapped against the same sheet and rejected, each on a single line: -9147459640944998301 (over distance, despite having a mushroom island and a quad monument), -4217104969789286797 (no second ocean monument, despite quad witch huts at spawn), -1056083236240939372 (crossroads too far, three connections, two monuments) and 6900926224957998486 (bastion too far).
The pattern across all four is worth more than the seeds themselves. In three of the four cases the seed passed the most valuable single criterion and was still turned down. Quad witch huts do not excuse a missing monument. A mushroom island does not rescue a cluster that sits past the distance budget. The checks are not a score that adds up — they are gates, and a seed has to clear all of them.
Those four have their own page, with the full coordinate set behind each decision: the rejected island seeds. A rejection is a server-grade judgement, not a verdict on whether a seed is playable — read that page before you write any of the four off for solo play.
How to run the same check on any island seed
You do not need a scanner. You need a creative copy of the world and about twenty minutes.
- Fly the perimeter out to 1,000 blocks first. Before you look at anything else. If the structures you care about sit outside that ring, you are done — do not let a mushroom island talk you out of it.
- Count the witch huts inside 256 blocks. Four is the target for a shared world. Two makes a decent solo seed. Zero means it is not a server seed.
- Count the monuments. One gets cleared once; two or four get farmed. Write the coordinates of the pair down — a dual monument 1,500 blocks out is worth recording but does not count as reachable.
- Travel to the bastion once. Actually make the trip. This is the line that rejected
6900926224957998486, and it is the check people skip because the other structures look good.
- Only then look at the nice-to-haves — mushroom island, warm ocean, coral. Those decide which passing seed you enjoy most. They never rescue a failing one.
If you play Bedrock, run the same list on Bedrock. The two editions do not share world generation, so a Java mapping tells you nothing about the code you are about to load. The Java vs Bedrock comparison shows what one code produces on each edition.
First day on a shared island spawn
The sheet above is for choosing a seed. This is for the first hour after a group spawns on one, and it applies whether the seed passed the review or failed it.
- Chop trees before anyone builds anything. Wood is the binding constraint on an island, not space. The rule from testing is that more trees beats a bigger island — a small island with twelve trees gives a group more options than a large one with three.
- Split the first hour into two jobs, not eight. One player works the coastline for shipwrecks and ruins; one player takes the wood and gets a crafting station and beds up. Eight players each doing their own thing on day one produces eight half-finished projects and no food.
- Mark the four hut coordinates on day one, even if the farm comes later. Structures get lost. Write them in chat and put them on a map before anyone wanders.
- Solve food before you solve redstone. The honest note from testing: you cannot sleep on the ocean, and supplies are difficult to restock. If the island has not produced a food source by the end of the first day, the better play is to get obsidian and open the Nether route rather than trying to make the island self-sufficient.
- Do not build the main base on the spawn island. On a server-grade seed the spawn island is where you stage, not where you live. The base goes where the structures are.
For spawns that were loaded and played specifically on Bedrock, with the supply situation noted for each, see the Bedrock island seeds page. For codes recorded against much older terrain generators, there is the 2011–2019 seed archive — read the version note on that page first.
Frequently asked questions
What makes an island seed good for a multiplayer server?
Structure density a group can use from one base: four witch huts inside a 256-block radius, more than one ocean monument, a bastion inside practical reach, warm ocean access, and all of it close enough that nobody travels 1,000+ blocks to reach any of it. A solo seed is judged on whether you survive the first night; a server seed is judged on whether several farms can run at once from the same spot.
Do I need quad witch huts if I play solo?
No. Quad witch huts matter because one farm can cover all four spawning areas, which is a throughput problem that only appears when several players are drawing from the same farm. Solo, two huts is plenty, and a seed with one hut and good wood supply is a better week-one experience than a seed with four huts and no trees.
What do you do if your survival island has no trees?
Work outward in a fixed order: any ruined portal chest first, then shipwrecks for planks and a treasure map, then ocean ruins. On one tested Bedrock seed the ocean ruins generated on land instead of underwater and held a stone axe, which removes the problem entirely. If none of that lands, get obsidian and open the Nether route rather than trying to make a treeless island self-sufficient.
Is a bigger island better than one with more trees?
No — more trees beats a bigger island. Wood is the constraint on any island spawn and space almost never is, so a small island with twelve trees gives you more options than a large one with three.
Do these server standards apply on Bedrock too?
The standards do, the seed codes do not. All three seeds on this page were mapped on Java 1.18 and above, and Java and Bedrock do not share world generation — the same code produces a different world on each edition. Load the world on the edition you actually play and re-run the four checks there.
Why is one seed on this page missing its spawn coordinate?
Because the tester did not write it down. Two of the three accepted seeds are recorded as structure coordinates only, with no spawn point in the log, so their distances are logged relative to the mapped area rather than to a recorded spawn. That gap is in the source and is left visible on purpose.
How do I check an island seed against these standards myself?
Load the world on Java 1.18 or above and run the four checks in order: are there four witch huts inside a 256-block radius, is there more than one ocean monument, is a bastion within practical reach, and do you have warm ocean access. Failing any one of them fails the seed, which is exactly why a seed can pass the most valuable single criterion and still be turned down. Every check that a seed failed is written on its row rather than averaged into a score.
What does "quad witch huts" actually mean?
Four witch huts close enough together that one farm covers all four spawning areas — the radius our testers use is 256 blocks. It is a throughput requirement, not a rarity requirement: three huts in a cluster give you three spawning areas, and four give you a farm that can supply several players at once from a single build. Solo, the number stops mattering, because two huts is already more than one player can drain.
What makes an island seed "server-grade" instead of just "good"?
A server-grade seed has to pass four checks at the same time: a witch hut within 256 blocks, an ocean monument within 1000 blocks, a pillager outpost within 600 blocks, and a second warm ocean or a second monument in another direction. A "good" seed only needs two or three of these. The reason all four are required is that a multiplayer server is four players spending the same first night, and any single missing structure forces a longer walk for the player who lands in its quadrant -- that is why the standard is on the /server-grade-island-seeds/ page rather than the main /survival-island-seeds/ page.
How many distance numbers does this site actually publish?
Eleven distinct ones, across forty-three threshold statements in fifteen of our thirty-three pages: 10, 11, 27, 200, 256, 300, 388, 500, 600, 1000 and 3000 blocks. Four of them — 200, 256, 600 and 1000 — are budgets you can apply to a seed you have not loaded yet, and three of those four are on this page. The rest describe one record once. We counted them all on the speedrun seeds page, because a budget you cannot find is a budget you cannot use.
Is three thousand blocks too far for a server-grade island?
Not on its own. The quad ocean monument on this page sits at -3100, -1100, which is over three thousand blocks from spawn in the overworld, and it survived review because Nether travel compresses the trip to about 388 blocks. What kills a seed is a whole cluster sitting past the budget with no shortcut to it, which is exactly what happened to the seed that was rejected for being slightly over a thousand blocks out.
Can I use the server-grade budget for a speedrun?
No, and the difference is in what each one asks. This budget asks whether anything is too far, because four players share one first night and the worst case has to be survivable. A speedrun budget asks whether everything is close enough, in the order you need it, on a clock. Every number here answers the first question, and none of them answer the second.
Want to run one of these with friends?
This whole sheet only matters if more than one person is going to use the world. If you want to host a Minecraft server and play these islands together — Java or Bedrock — Shockbyte is a low-cost host that sets up in minutes.
Use code LAUNCH for 25% off your first month. If you sign up through that link, I may receive a commission at no extra cost to you. It does not change what you pay.
