Skip to main content

Creating Images

Updated over 2 weeks ago

The Image feature is the perfect tool for the job. By using a single reference image as a style guide along with your text prompt, you can generate draft images to create models from.

How It Works

This feature analyzes the visual style of an uploaded reference image. It looks at the colors, textures, lighting, and overall aesthetic. When you generate a new 3D model, the AI will apply those same characteristics to the new object, creating a seamless match.

Step-by-Step Guide

  1. Prepare Your Reference Image: Start with a high-quality 2D image that represents the style you want to achieve. This could be an image of a model you've already made, a piece of concept art, or a picture with a specific aesthetic. A clear, well-lit image works best. Note: you can also just use the text prompt to generate an image.

  2. Upload Your Image: Drag and drop or upload your prepared reference image into the designated area:

  3. Add a Text Prompt: In the text prompt box, describe the new object you want to create. Be specific about what you're making, but let the reference image guide the style. For example, if your reference image is a stylized tree, you could use a prompt like "a stylized mushroom" or "a matching rock."

  4. From there, you can select your AI model type (Nano Banana, Flux Kontext or Flux Dev), image strength, and aspect ratio:

    Please note that image strength can only be applied to Flux Dev.


    Choosing the right tool for your 3D image generation depends on your needs. Here's a breakdown to help you decide between our different models.

    For generating high-quality images of 3D structures, we recommend Flux Kontext. It's our best AI for this purpose, producing excellent results. If speed and cost are your top priorities, consider Flux Dev. It works faster and is more affordable, though the image quality may be lower. Flux also offers greater stylistic flexibility, making it an excellent choice for generating images in styles like oil painting or conceptual art.

    The Nano Banana model is specifically designed to create images that are ideal for 3D generation. Its drawings feature a clearer structure and more accurate lighting and shadows. This makes it a great option if you plan to use the generated image as a blueprint for a 3D model.

  5. Generate and Refine: Click "Generate" and you will get four model drafts based on your prompt/reference image:

  6. From there, you can expand the image (magnifying glass in right hand corner), load it into the Image to 3D prompt (arrow at bottom), or generate the model (Image to 3d button with shows credit cost):

This tool is a great way to maintain a cohesive style across all your 3D assets, by consistently using the same reference image for all your creations.

Did this answer your question?