SNNIB: Spiking Neural Networks Into Blender

:local:
:depth: 3

Note

this add-on consists of two parts: the blender add-on; a python package used create files compatible with SNNIB;

• if you find this add-on useful in your work, an acknowledgement would be appreciated:

@software{PY_Steinwender2026_snnib,
	author    = {{Steinwender}, Lukas},
	title     = {SNNIB: Spiking Neural Networks Into Blender},
	month     = Mar,
	year      = 2026,
	version   = {latest},
	url       = {https://github.com/TheRedElement/snnib.git}
}

Example Renders

randomly generated network	imported brian2 network (source)
imported brian2 network (400 neurons, 1260 synapses, source)

Note

you can also render much larger networks depending on your hardware. on a 16GB RAM, 16 core laptop I tested up to 1600 neurons. The main issue you will run into is that there is so much going on, that it is hard to distinguish individual neurons and neurites.

Quickstart

more detailed documentation can be found in the readthedocs page

Add-on

Installation

1. download releases/snnib.zip
2. in blender
   1. navigate to Edit > Preferences > Add-ons
   2. drag and drop the downloaded file (snnib.zip) into the window
   3. click OK

Navigation

see Video Tutorials for screen recordings

1. in Viewport Display expand the right side panel (by hitting n)
2. initialization (only needs to be done once right after loading)
   1. run all the operators in the Actions > Initialization box in order
      1. done by clicking the respective buttons
      2. necessary to have all building blocks for a functioning network available
3. generate a Template Neuron
   1. will be used as source to instance neurons in your network
   2. it is recommended to use the generated template neuron and adjust it to your liking via the associated Geometry Nodes modifier
      1. you just need to modify the Geometry Nodes on the template neuron as all changes will be reflected on all instances
4. adjust the Settings
   1. provide a Network Container
      1. I recommend a cuboid because the add-on uses the objects bounding box to define the neuron positions
   2. adjust the remaining settings or styling in the Template Neurons Geometry Nodes
5. make sure to Apply Scale on the Network Container (and potentially also the Template Neuron)
6. hit Build SNN to generate your network

Mappings

• one time-step ($dt$) in a SNN simulation is mapped to a single frame in blender

Python package

Installation

• simply install via pip

pip3 install git+https://github.com/TheRedElement/snnib.git

for tutorials checkout tutorials

Currently supported simulators

• random network generation
• brian2

For developers

Compiling the add-on

If you want to compile the add-on yourself (i.e., in case you made some changes to a forked repo and want to compile an updated version) you can do so by calling the following from the repository root:

bash publish.sh

Known Restrictions

• because a lot of geometry is generated when building a large network, crtl + z will likely fail

TODO:

• geo nodes node trees and shader nodes node trees do not persist when reloading the .blend file
• save function (to store network randomly generated with snnib)

Video Tutorials

generating random network loading network from file control using geometry nodes