- PC
- Mobile
The generation loop
Write a prompt
Describe what you want. Subject, action, setting, style. The more specific, the more predictable the result. See Prompting Basics.
Choose a model
Each model has a distinct visual style. Realism for photographic output, Anime for illustrated, Cinematic for film-grade lighting, and so on. See Models.
Add modifiers (optional)
Layer on a LoRA, a Style pack, or an Action to push the output in a specific direction. These stack on top of the model — they don’t replace it.
Set resolution (optional)
The default resolution works for most use cases. Change it if you need a specific aspect ratio or plan to animate the result. See Parameters.
What affects the output
| Setting | What it does | Where to learn more |
|---|---|---|
| Model | Sets the overall visual style | Models |
| LoRA | Adds a fine-tuned layer on top of the model | LoRAs |
| Style | Applies a curated aesthetic pack | Styles |
| Action | Applies a predefined content modifier | Actions |
| Preset | Saves and reloads a full combination of settings | Presets |
| Resolution | Controls output dimensions and aspect ratio | Parameters |
| Seed | Locks the random starting point for reproducibility | Parameters |
Free tier
Free accounts can generate five images per day using Realism and Anime. All other models, LoRAs, Styles, Actions, InPaint, and Private Generation require Premium. See /premium.From image to video
Any image you generate can be used as the first frame of a video. This is the primary path into video generation — generate a strong first frame here, then animate it. See Video: Workflow.Models
Pick the right model for your style.
Prompting Basics
Structure a prompt that gets consistent results.
InPaint
Edit a specific area of any generated image.
Video: Workflow
Animate your image into a video.