Image Sizes¶

OmniImage provides standard size constants for common image dimensions.

Size Constants¶

Constant	Dimensions	Aspect Ratio	Use Case
`Size256x256`	256x256	1:1	Thumbnails, icons
`Size512x512`	512x512	1:1	Small images, previews
`Size768x1024`	768x1024	3:4	Portrait photos
`Size1024x768`	1024x768	4:3	Landscape photos
`Size1024x1024`	1024x1024	1:1	Standard, social media
`Size1024x1792`	1024x1792	9:16	Tall portrait, mobile
`Size1792x1024`	1792x1024	16:9	Wide landscape, banners

Provider Support¶

Size	OpenAI	Fal AI
256x256	DALL-E 2	No
512x512	DALL-E 2	Yes
768x1024	No	Yes
1024x768	No	Yes
1024x1024	All	Yes
1024x1792	GPT Image, DALL-E 3	No
1792x1024	GPT Image, DALL-E 3	No

Usage¶

import "github.com/plexusone/omniimage/provider"

// Square image
resp, _ := client.Generate(ctx, &provider.GenerateRequest{
    Model:  omniimage.ModelGPTImage2,
    Prompt: "A profile picture",
    Size:   provider.Size1024x1024,
})

// Portrait
resp, _ := client.Generate(ctx, &provider.GenerateRequest{
    Model:  omniimage.ModelGPTImage2,
    Prompt: "A full-body portrait",
    Size:   provider.Size1024x1792,
})

// Landscape
resp, _ := client.Generate(ctx, &provider.GenerateRequest{
    Model:  omniimage.ModelGPTImage2,
    Prompt: "A panoramic view",
    Size:   provider.Size1792x1024,
})

Custom Sizes (Fal AI)¶

Fal AI supports arbitrary dimensions via ProviderOptions:

resp, _ := client.Generate(ctx, &provider.GenerateRequest{
    Model:  omniimage.ModelFluxPro,
    Prompt: "A website banner",
    ProviderOptions: map[string]any{
        "image_size": map[string]int{
            "width":  1920,
            "height": 480,
        },
    },
})

Choosing the Right Size¶

Content Type	Recommended Size
Profile pictures	1024x1024
Social media posts	1024x1024
Mobile wallpapers	1024x1792
Desktop wallpapers	1792x1024
Thumbnails	256x256 or 512x512
Product images	1024x1024
Banners	Custom via Fal AI

Size and Quality¶

Larger sizes generally provide:

More detail
Better for printing
Larger file sizes
Longer generation time

Performance

For rapid prototyping, use smaller sizes (512x512). Scale up to production sizes once you've refined your prompt.