[!WARNING] Warning, uses experimental package
comfy-envto attempt a one click isolated install. Will download and use pixi package manager.
Three options, in order of speed → reliability:
Pixal3D in the Manager and click Install from the highest version displayed. If that doesn't work, try nightly.https://github.com/PozzettiAndrea/ComfyUI-Pixal3D.git.cd ComfyUI/custom_nodes
git clone https://github.com/PozzettiAndrea/ComfyUI-Pixal3D.git
cd ComfyUI-Pixal3D
pip install -r requirements.txt --upgrade
python install.py
Please report any problems you hit during installation or use of my nodes — open a Discussion or Issue. Very grateful for your help! 🙏
ComfyUI nodes for Pixal3D (SIGGRAPH 2026, TencentARC) — pixel-aligned image-to-3D generation. Single image in, textured GLB out.
| Node | Purpose |
|---|---|
Pixal3DLoadPipeline | Loads the cascade pipeline + four DinoV3 cond models. Auto-downloads Pixal3D weights. |
Pixal3DLoadMoGe | Loads MoGe-2 for camera-intrinsic estimation. |
Pixal3DPreprocessImage | Background removal + alpha bbox crop + 1024-max resize. |
Pixal3DEstimateCamera | Runs MoGe-2 to estimate camera_angle_x and distance from the input image. |
Pixal3DGenerate | Runs the four-stage cascade (SS → shape LR 512 → shape HR 1024 → texture 1024). |
Pixal3DExtractGLB | Extracts a textured GLB via o_voxel.postprocess.to_glb. Saves to output/. |
1024_cascade with low_vram=True. More for 1536_cascade.Questions or feature requests? Open a Discussion on GitHub.
Join the Comfy3D Discord for help, updates, and chat about 3D workflows in ComfyUI.
Built on the work of the Pixal3D authors (Li, Zhao, Chen, Hu, Guo, Zhang, Shan, Hu — Tsinghua / Tencent ARC / Victoria University of Wellington), Microsoft TRELLIS.2, Direct3D-S2, MoGe (Microsoft), and DINOv3 (Meta).
Built with DINOv3.
Wrapper authored by Andrea Pozzetti.
Contributions are welcome! Please feel free to submit issues and pull requests.
64 commits
Python
97.4%
HTML
1.5%
JavaScript
1.1%
[!WARNING] Warning, uses experimental package
comfy-envto attempt a one click isolated install. Will download and use pixi package manager.
Three options, in order of speed → reliability:
Pixal3D in the Manager and click Install from the highest version displayed. If that doesn't work, try nightly.https://github.com/PozzettiAndrea/ComfyUI-Pixal3D.git.cd ComfyUI/custom_nodes
git clone https://github.com/PozzettiAndrea/ComfyUI-Pixal3D.git
cd ComfyUI-Pixal3D
pip install -r requirements.txt --upgrade
python install.py
Please report any problems you hit during installation or use of my nodes — open a Discussion or Issue. Very grateful for your help! 🙏
ComfyUI nodes for Pixal3D (SIGGRAPH 2026, TencentARC) — pixel-aligned image-to-3D generation. Single image in, textured GLB out.
| Node | Purpose |
|---|---|
Pixal3DLoadPipeline | Loads the cascade pipeline + four DinoV3 cond models. Auto-downloads Pixal3D weights. |
Pixal3DLoadMoGe | Loads MoGe-2 for camera-intrinsic estimation. |
Pixal3DPreprocessImage | Background removal + alpha bbox crop + 1024-max resize. |
Pixal3DEstimateCamera | Runs MoGe-2 to estimate camera_angle_x and distance from the input image. |
Pixal3DGenerate | Runs the four-stage cascade (SS → shape LR 512 → shape HR 1024 → texture 1024). |
Pixal3DExtractGLB | Extracts a textured GLB via o_voxel.postprocess.to_glb. Saves to output/. |
1024_cascade with low_vram=True. More for 1536_cascade.Questions or feature requests? Open a Discussion on GitHub.
Join the Comfy3D Discord for help, updates, and chat about 3D workflows in ComfyUI.
Built on the work of the Pixal3D authors (Li, Zhao, Chen, Hu, Guo, Zhang, Shan, Hu — Tsinghua / Tencent ARC / Victoria University of Wellington), Microsoft TRELLIS.2, Direct3D-S2, MoGe (Microsoft), and DINOv3 (Meta).
Built with DINOv3.
Wrapper authored by Andrea Pozzetti.
Contributions are welcome! Please feel free to submit issues and pull requests.
64 commits
Python
97.4%
HTML
1.5%
JavaScript
1.1%