How It Works
Start with a clear idea
Use the prompt box when you already know what you want, or start from tags if you want a faster setup. The strongest results usually come from describing the subject, mood, framing, and style in one clean direction.
Shape the result with filters
Use the Body Type, Styles, and Character groups to steer the output. These tags help the model lock onto the look, number of people, pose, fantasy traits, and overall art direction without making the prompt messy.
Generate and review
When your setup looks right, generate the image and let the model render it. The result appears in the main viewport where you can review it, regenerate, download it, or reuse the tags for another variation.
Iterate with intention
If the first result misses, keep the subject centered, simplify conflicting ideas, and add only the details that matter most. A tighter prompt plus a few strong tags usually beats a long list of unrelated descriptors.