PlaneToPBR is a free Blender extension that generates a fully textured PBR plane from a single image in one click.
Watch the full demo video here: https://youtu.be/AtydKv722z8
The extension automatically generates and applies:
- Diffuse
- Normal
- Roughness
- Depth
- Prompt-based segmentation mask
This allows artists to quickly convert reference images into usable 3D textured geometry.
Compatible with Blender 4.2 and newer.
- Open Blender
- Go to Edit → Preferences
- Click Get Extensions
- Open the dropdown menu and select Install from Disk
- Select the
PlaneToPBR.zipfile
- Enable the extension after installation.
After installing, open the PlaneToPBR panel in the 3D viewport sidebar.
The panel is located in the 3D viewport sidebar.
Optionally enter a prompt to guide the AI segmentation.
Example:
Windows
Click the file browser button and select your source image.
Then choose the image file.
PlaneToPBR Pro / GPU generation is authenticated from the add-on preferences, not from the sidebar panel.
Open Edit -> Preferences -> Add-ons -> PlaneToPBR and log in there.
The standard Hugging Face generation remains free and does not require login.
Click either:
- Generate PBR (HuggingFace) for the free flow
- Generate PBR Pro (GPU) for the authenticated platform flow
The add-on will:
- Send the image to the AI pipeline
- Generate PBR maps
- Create a correctly scaled plane
- Build the material automatically
The extension automatically builds a full PBR shader network.
For best depth results, adjust modifiers:
- Subdivision
- Displacement
Example settings:
PlaneToPBR sends the selected image to a remote AI inference service hosted on Hugging Face in order to generate PBR texture maps.
Only the image selected by the user is transmitted.
No Blender project data or personal information is collected or stored.
PlaneToPBR relies on several open-source AI models:
| Model | Purpose | Repository |
|---|---|---|
| CLIPSeg | Text-guided segmentation (mask generation) | https://github.com/timojl/clipseg |
| DeepBump | Normal & roughness map generation | https://github.com/HugoTini/DeepBump |
| MiDaS | Depth estimation | https://github.com/isl-org/MiDaS |
GPL-3.0-or-later