This Blender add-on allows you to import 3D geometry, splats, or radiance field data generated using the Triangle Splatting for Real-Time Radiance Field Rendering method. Perhaps it can import other OFF files as well.

It is designed for artists, researchers, and developers who want to bring advanced neural rendering outputs directly into Blender for visualization, animation, or further processing. It works in Blender 4.4, perhaps in later or earlier versions too, but this has not been tested.

• Import triangle splatting output files into Blender, streaming large files.
• Preserve material and shading information where possible.
• Compatible with Blender's Cycles and Eevee renderers.
• Simple and streamlined UI in the Import menu.

1. Download this repository as a .zip file.
2. Open Blender and go to: Edit → Preferences → Add-ons → Install...
3. Select the .zip file and click Install Add-on.
4. Enable the add-on by checking the box next to its name.

1. Go to: File → Import → Triangle Splatting Object File Format (.off)
2. Select your exported file from the Triangle Splatting pipeline.
3. Adjust import settings as needed.
4. Click Import to load the scene into Blender.

This add-on is released under the MIT License. Please note that the Triangle Splatting method itself may have its own license and terms — check the original repository or paper for details.

This add-on builds upon the work presented in:

Held, Jan, Vandeghen, Renaud, Deliège, Adrien, Hamdi, Abdullah, Cioppa, Anthony, Giancola, Silvio, Vedaldi, Andrea, Ghanem, Bernard, Tagliasacchi, Andrea, & Van Droogenbroeck, Marc. Triangle Splatting for Real-Time Radiance Field Rendering. arXiv, 2025. If you use this add-on in research or publications, please cite the original Triangle Splatting paper:

@article{Held2025Triangle,
title   = {Triangle Splatting for Real-Time Radiance Field Rendering},
author  = {Held, Jan and Vandeghen, Renaud and Deliege, Adrien and Hamdi, Abdullah and Cioppa, Anthony and Giancola, Silvio and Vedaldi, Andrea and Ghanem, Bernard and Tagliasacchi, Andrea and Van Droogenbroeck, Marc},
journal = {arXiv},
year    = {2025},
}