Image Settings
Beyond the prompt, image settings affect composition, detail level, and quality.
Image Size & Aspect Ratio
The aspect ratio significantly affects composition. Choose wisely:
| Ratio | Best For | Examples |
|---|---|---|
| Portrait (e.g. 768x1024) | Standing characters, full body, vertical compositions | Pinups, character sheets, standing poses |
| Landscape (e.g. 1024x768) | Lying poses, feral characters, scenes with backgrounds | Feral art, scenic shots, multi-character |
| Square (e.g. 960x960) | Safe default, works for most prompts | Headshots, general compositions |
Size Limits
- Any size with a width or height of 1200px or more is limited to 1 image per request
- Upscaling is not available for very large sizes
- Larger sizes require more processing time
Number of Samples
You can generate 1 to 3 images per request (depending on your plan). Generating multiple samples is useful to pick the best result - AI generation has an element of randomness, and each sample is unique.
Image Reference
You can paste an image URL to guide the generation. The AI will use it as a reference for composition and design.
- Useful for recreating a pose from a reference
- Helps preserve character design across generations
- Works well with sketches as input
- Does not copy the artistic style - only the structure and composition
- Your text prompt still works alongside the reference - use it to specify details the reference doesn't cover