A utility script made in C and Python to find seeds that match biome and size requirements. This project contains functionality that’s already in the website mcseeder. The problem is: actual useful seed requirements take ages to process (and the GUI version doesn’t support MacOS). This is a (multithreaded) script which you can host yourself (or just run on your computer).
I originally made the script for my server, BingoMC. I need multiple different seeds with at least 8 specific biomes in a radius of 2500 blocks for the game to work correctly (otherwise it would be impossible). I store them (pregenerated) in an S3 bucket and download them onto the server once a game boots up.
The script works by generating partials of a seed map (all credits go to https://github.com/Cubitect/cubiomes 😛) and taking samples to see if the biomes match. Seeds will be tried incrementally (1, 2, 3, 4…) and checked to see if it contains all biomes by trying a grid of single coordinates with an amount blocks between them (15 by default).
The script takes some options: the seed range, block radius, step size, and the amount of threads. These settings all dictate how long the script will take. I found that 30 seeds in a radius of 2500 blocks, with a step size of 15, on one thread will take about 1 minute, which I thought was surprisingly fast.
| Setting | Default | Description |
|---|---|---|
| Seed range | 0-200 | The actual range of seeds to match against. More seeds means more processing. |
| Block radius | 2500 | The radius of blocks to look for from X=0 and Z=0. 2500 will be a 5000x5000 area. A larger radius means more processing. |
| Step size | 15 | The amount of blocks that will be in between a sample. A larger number means less processing, but also less accuracy. It’s recommended to keep this below ±35, as you can miss biomes (e.g. rivers) with such a large step size |
| Thread count | 4 | The amount of threads used to divide the workload of the seed range. 200 seeds and 4 threads means each thread does 200 / 4 = 50 seeds. More threads will make things faster, but you can’t set more threads than your computer has logical cores (of course) |
Clone this repo:
git clone --recursive [email protected]:jensjeflensje/minecraft_seed_finder.gitThe C program has a fixed list of biomes it wants. You can change this.
To compile the program, execute the compile.sh script.
When it's successfully compiled, you're able to change the settings (capital letters) in the check_seeds.py script, and execute it.
You'll see a file named seeds.txt be created with the found seeds.