Create optimal part layouts for CSM’s Image to 3D using ChatGPT’s updated Text-to-Image generation:
In ChatGPT
Open ChatGPT, ensuring you're using the updated Model 4o.
Clearly describe your object parts laid out on a sheet suitable for 3D generation. Useful keywords include:
…with all parts laid on a sheet for Image to 3D.
…create all major parts on a sheet for Image to 3D-based assembly, no spurious parts, only major components.
Specify parts or angles clearly if initial results aren't ideal:
Request clearer layouts or angled parts for better depth.
Adjust style if needed (e.g., more cartoon-like or thematic elements).
When satisfied, download the generated image.
In CSM
On your homepage, click Parts & Asset Pack.
Drag & drop or upload your downloaded image.
Use segmentation tools on the left to select each piece separately. Click Mask Image after each selection. When done, click Next.
Configure mesh generation settings. Refer to our Image to 3D settings guide here for assistance.
Generated meshes appear grouped under Asset Packs.
Optionally, retexture your meshes:
On each mesh page, click Retexture.
For detailed instructions, refer to our retexturing tutorial.
Download each mesh file using your preferred format.
In Blender
Import all your mesh parts into Blender.
Scale and assemble your object as needed.
Your fully assembled mesh is now ready.
For inspiration, watch our rendered assembly demo on our YouTube channel.