Image

Generate or edit an image using a diffusion model

returns GeneratedImage

Generates an entirely new image from a text prompt using models like DALL·E, Gemini or Imagen. This is the primary entry point for creative image generation.

Generating an Image

Texture2D result = await "A cat surfing a wave"
    .GENImage()
    .SetModel(ImageModel.DallE3)
    .ExecuteAsync();

DALL·E-specific Configurations

Texture2D result = await "A cyberpunk city at night"
    .GENImage()
    .SetModel(ImageModel.DallE3)
    .SetSize(ImageSize._1024x1024)
    .SetQuality(ImageQuality.HD)
    .SetStyle(ImageStyle.Vivid)
    .ExecuteAsync();

Google-specific Configurations (Gemini, Imagen3)

Texture2D result = await "An astronaut riding a horse"
    .GENImage()
    .SetModel(GoogleModel.Gemini2_0_Flash_Exp_Image_Generation)
    .SetAspectRatio(AspectRatio.Square)
    .SetPersonGeneration(PersonGeneration.Unspecified)
    .ExecuteAsync();

Results are saved to a temporary location unless you call .SetOutputPath().

---

GENInpaint

returns GeneratedImage

Edits an existing image using a prompt. Ideal for adding, removing, or changing parts of an image.

Basic Usage

// The given image is used as a base, and the model performs localized edits.

Texture2D result = await myTexture
    .GENInpaint("Add a crown on the cat’s head")
    .SetModel(ImageModel.DallE2)
    .ExecuteAsync();
    
// You can provide a mask using InpaintPrompt if you need precise region control.

DALL·E 3 does not supports inpaint.

---

GENVariation

returns GeneratedImage

Creates new stylistic variants of an existing image. This is useful for generating concept variations or alternatives.

Only DALL·E 2 currently supports image variation.

Basic Usage

Texture2D result = await myTexture
    .GENImageVariation()
    .SetModel(OpenAIModel.DallE2)
    .ExecuteAsync();