Cyber Draw — a small Android arcade game I built with py4godot, Godot, and ONNX Runtime. Draw a symbol to fire off an explosion, a shockwave, or a laser. Play it on itch.io

Penpot's Whac-A-Mappole by Carlos Padial — a real-time interactive video mapping installation, projected onto a wall of cardboard reliefs and built entirely with free software. It uses a tiny bit of py4godot: a Python script inside Godot reads raw USB data from four mice used as game controllers. The rest of the project is far more impressive than that one detail. Well worth a read! Read the write-up

There are a few other cool projects bringing Python to Godot. Check them out!

• godot-python-extension
• godot-python

Python scripting for Godot. This GDExtension plugin allows you to use Python like GDScript, accessing the Godot engine's features and benefiting from Python's vast ecosystem.

Inspiration for this project was taken from: https://github.com/touilleMan/godot-python.

This project is currently in an early phase of development. Consider this more as a demo and don't use this for bigger projects. At the moment only Windows 64 bit and Linux 64 bit are supported.

Examples for this plugin can be found here: Examples

If you want a small introduction into py4godot, check out the developer guide: Developer Guide

• Windows 32 bit
• Windows 64 bit
• Linux 32 bit
• Linux 64 bit
• macOS
• Android
• iOS

Download the .zip from Releases, extract it and copy the folder into the addons folder on the root of your project.

You need a Python 3 version of at least 3.11 with pip installed.

As this project depends on some modules which are downloaded by pip, if you don't want to mess with your packages, you should set up a virtual enviroment. The following code should be executed after cloning the repo

$ python -m venv virtual_python #creating the folder with the virtual environemnt
$ .\virtual_python\Scripts\activate #starting the virtual enviromnent 
$ pip install -r requirements.txt #load dependencies from the textfile requirements.txt

We want to generate the files we will later use for the cython compilation like this:

$ python generate.py

Now we use Cython to generate the C++ files for our Python classes.

$ python cythonize_files.py

if you want faster building, you can build a minimized version. This doesn't include all the classes, but should be enoug for most development and simple projects. For this just call generate.pyand cythonize_files.py with the following options:

$ python generate.py -dev_build=True
$ python cythonize_files.py -mode="dev"

Now, we just need to compile the project. The result of the compilation can be found on build/py4godot

$ python build.py --target_platform=windows64 --compiler=msvc 

Building for linux was tested on Ubuntu 20.04.4 LTS

The following code should be executed after cloning the repo

$ python3 -m venv venv  # creating the folder with the virtual environment
$ source venv/bin/activate  # starting the virtual environment
$ pip install -r requirements.txt  # load dependencies from the text file requirements.txt

If you are on Python 3.12, you unfortunately have to use this workaround:

$ pip install setuptools #Weirdly cython on linux depends on distutils. But python3.12 doesn't provide this anymore. So we have to use this workaround
$ python copy_distutils.py

We want to generate the files we will later use for the cython compilation like this:

$ python generate.py

Now we use Cython to generate the C++ files for our Python classes.

$ python cythonize_files.py

if you want faster building, you can build a minimized version. This doesn't include all the classes, but should be enoug for most development and simple projects. For this just call generate.pyand cythonize_files.py with the following options:

$ python generate.py -dev_build=True
$ python cythonize_files.py -mode="dev"

Now, we just need to compile the project. The result of the compilation can be found on build/py4godot

$ python build.py --target_platform=linux64 --compiler=gcc 

Here you can see a basic example of how this project can be used. Please note, that the classname you define under @gdclass must match the filename

# file: node3d.py
from py4godot.methods import private
from py4godot.signals import signal, SignalArg
from py4godot.classes import gdclass
from py4godot.classes.core import Vector3
from py4godot.classes.Node3D import Node3D

@gdclass
class node3d(Node3D):

	# define properties like this
	test_int: int = 5
	test_float: float = 5.2
	test_bool: bool = True
	test_vector: Vector3 = Vector3.new3(1,2,3)

	# define signals like this
	test_signal = signal([SignalArg("test_arg", int)])


	def _ready(self) -> None:
		pass
		# put initialization code here

	def _process(self, delta:float) -> None:
		pass
		# put dynamic code here

	# Hide the method in the godot editor
	@private
	def test_method(self):
		pass