When the user clicks Visualize, the system generates images where:
the same character stays consistent (face, hair, age, vibe, outfit rules)
the environment matches the selected world/setting (volcano vs forest)
the style stays consistent (cinematic, watercolor, anime, etc.)
props/lore rules are respected (weapons, magic glow, era, tech level)
The render pipeline shouldn’t rely on a single prompt typed by the user. It should build a structured prompt from cards.
Think of it as:
Visualize Prompt =
Character Anchor (immutable identity)
World Anchor (rules + aesthetics)
Location/Setting Module (forest / volcano / tavern)
Scene Moment (what’s happening right now)
Camera + lighting (shot language)
Negative constraints (what to avoid)
Please authenticate to join the conversation.
Proposed
💡 Feature Request
11 days ago

Manon Doucet
Get notified by email when there are changes.
Proposed
💡 Feature Request
11 days ago

Manon Doucet
Get notified by email when there are changes.