How I generated a high quality 3D model in Blender, from a simple 2D image, using Claude 3.7! Screen capture + link + step-by-step process below👇

First, you'll need Blender + a Blender MCP. For this project, I used @VisionaryxAI 's MCP available on Github. See The install steps are simple and explained on the Github page.

Feed an image to Claude (or Cursor) - I selected a simple isometric scene initially made with a custom model on @Scenario_gg . The custom model is "Blocky Elements" and typically creates minimalist 3D render-style images ➡️ a relevant model to use for this case.

Ask Claude to describe it to Blender (MCP) with a simple prompt. You can try "Recognize any elements of this scene individually. Describe them precisely, then build them in a scene." or anything similar. Ask for refinements if needed.

Eventually you should get a 3D model in Blender with elements of your image. Now, it might not be perfect (it wasn't in this case) and I'll be honest... I have no idea - yet - why some elements were "disjointed“ and others weren't.

Anyway at this stage you can either prompt for modifications (it sometimes works surprisingly well) or if you feel for it and it's faster, just move some elements around with the Blender gizmos.

After just a few minutes, you'll have a 3D version of your image ready to edit further in Blender, export, or use however you need. 🚀🚀

Can Machine Learning beat the market? Check out this post on my free Substack where I share code and commentary for an XGBoost model and a Random Forest model that both deliver powerful performances.

Cool!!

Amazing! Thanks for sharing

Wow!