Skip to main content

Generation Basics

Overview​

Generation Basics explains how to control the results produced by the AI. These controls are organized into sections in the main panel and allow you to guide the content, structure, style, and quality of your output.

Available controls may vary depending on the selected model.

Basic AI Settings​

This section contains the main controls used to transform or edit your image.

  • Prompt Describe what you want to create or change. Start simple, then add details to refine your result.

  • Presets Apply predefined styles to quickly explore different looks. Presets can be combined with your prompt. Effects may vary depending on your Text Guidance settings and the selected model.

  • Seed Controls variation in your results. Changing the seed produces different outputs, while keeping it the same helps reproduce a similar result. Useful when generating variations or comparing results.

    To generate multiple variations at once, enable Image Variants in AI Engine > Settings.

Advanced AI Settings​

This section provides more control over how the AI interprets your input.

  • Negative Prompt (if supported by the model) Describe what you want to avoid in the result.

  • Text Guidance (if supported by the model) Controls how strongly the result follows your prompt. Higher values follow the prompt more closely, while lower values allow more variation.

  • Steps Controls how much processing is used to generate the image. More steps can improve quality, but take more time.

AI Processing Resolution​

AI Processing Resolution controls the resolution used by the AI while transforming or editing your source image.

Higher processing resolutions can preserve more detail and may improve results, especially for complex images. They can also take more time and require more memory.

If generation is slow or fails because of memory limits, try using a lower AI Processing Resolution or a model variant better suited to your hardware.

Image Transformation Settings​

Image Transformation Settings control how Vision FX uses your source image during generation.

Transformation Strength (if supported by the model)​

Transformation Strength controls how much the AI is allowed to change the source image.

  • Lower values preserve more of the original image.
  • Higher values allow stronger changes and more creative reinterpretation.

Use lower strength when you want a subtle edit or style change. Use higher strength when you want the result to move further away from the original image.

Structure Guidance (if supported by the model)​

Structure Guidance helps preserve the composition of your source image while transforming its appearance.

When enabled, Vision FX can use the source image to guide important structural elements such as shapes, outlines, character poses, object placement, architectural layouts, and overall scene composition.

Use Structure Guidance when you want to:

  • Keep the same general layout while changing the style
  • Preserve a pose, object arrangement, or architectural structure
  • Create more predictable transformations from a source image
  • Explore new looks without losing the original composition

Structure Guidance is available with supported models.

Style Reference (if supported by the model)​

Style Reference lets you use another image to guide the visual style of your generated result.

When enabled, Vision FX uses the reference image alongside your prompt to help influence the look and feel of the output. This can include the overall aesthetic direction, color palette, lighting, mood, texture, or artistic style.

Use Style Reference when you want to:

  • Create images with a look inspired by another image
  • Keep a more consistent visual direction across multiple generations
  • Combine your prompt with a specific style, mood, or visual treatment

Your prompt still defines what you want to create or change. The reference image helps guide how the result looks.

Style Reference is available with supported models.

Output Settings​

Output Settings provide built-in tools for refining or preparing generated images.

  • AI Image Enhancement: Improve generated images with refinement modes for denoising, sharpening, smoothing, and overall cleanup.
  • Smart Background Removal: Remove the background from the image.
  • Smart Crop: Automatically focus on the main subject.

AI Image Enhancement can use either your CPU or GPU when supported, giving you more control over performance on your system.

Generate​

Click Generate to create or update your result.

  • Generation transforms the current source image using your prompt and settings.
  • You can adjust settings and generate again to refine your result.

Tips​

  • Start simple and refine your prompt gradually.
  • Use presets to explore styles quickly.
  • Use Transformation Strength to control how much the source image changes.
  • Use Structure Guidance when you want to preserve layout, pose, or composition.
  • Use Style Reference when you want to guide the overall look using another image.
  • Adjust one setting at a time to understand its effect.
  • Duplicate images to compare how different settings affect results side by side.
  • Generate multiple times to improve your results.