Product

Text to 3D with BlueFox

Describe an asset, approve a picture of it, and BlueFox, Hydrilla's 3D model, builds it as a segmented 3D model in Hydrilla AI. About BlueFox

The 3D character generator panel with a typed prompt for a greyhound explorer in a long coat.

1Prompt

The reference image rendered from the prompt: a greyhound character in a long coat and scarf.

2Reference image

The finished 3D model of the character, shown untextured and textured side by side.

33D model

Text to 3D is the right start when the idea exists only as words: a prop list from a level designer, a line in a script, or a note from an art director. You write what the asset is and how it will be used, and Hydrilla turns that description into a model you can preview and take further.

Four steps from prompt to model

Text to 3D starts with a reference image that you can refine before the 3D model is generated. That picture is your chance to agree on the look before any geometry exists.

  1. Write a prompt naming the subject, its style, and its use.
  2. An image model renders a reference, at Low or High quality, in a square, landscape, or portrait frame.
  3. Adjust the image with an Edit prompt until it looks right.
  4. Generate the 3D model, then preview it and export.

Every model comes back UV-unwrapped, with PBR maps, so it is ready for material work from the start. What BlueFox returns

Write prompts that say how the asset will be used

Be specific about scale and use. A prompt that names the use, the scale, and how close the camera will get gives BlueFox a clearer brief than a list of adjectives. The examples below show the difference that one extra detail makes.

PromptWhy it works
Low-poly wooden crate for a Unity mobile gameNames the style and engine
Weathered brass ship lantern with worn edgesSets the material and detail level
Walnut lounge chair at real scale, for interior rendersGives scale and use
Cargo container for a background setSays it sits far from camera

Refine the image before you spend on 3D

Settle the look at the image stage, where a change is one more Edit prompt. Adjust color, proportion, or wear on the picture until it reads the way you want, then generate the model once. It keeps review short and credits focused on the version you actually approve.

When another start fits better

If you already have a photo or concept image, start from it with image to 3D. If the job is a whole set of pieces rather than one model, brief Water Agent instead. See what a model costs