Scene Illustrations¶
Visions of Adventure¶
The Visions tab (4) lets you generate AI artwork depicting dramatic moments from your session.
How It Works¶
Image generation is a two-step process:
- Scene Crafting — an LLM reads your transcript and crafts a vivid scene description focusing on one dramatic moment
- Image Generation — the description is used to create an image directly on your Mac, with no internet connection required
Generating an Image¶
- Switch to the Visions tab
- Click Generate Image
- Watch the progress phases:
- "Crafting scene..." — AI is writing the scene description
- "Generating image..." — mflux is rendering the image locally
- Done! — image appears in the gallery
Hidden Feature: Edit the Scene Description
Before generation begins, you can edit the scene description that the AI crafted. This lets you steer the image toward a specific moment or adjust details.
Hidden Feature: Character Appearance Consistency
If your character roster has visual descriptions, they're included in the scene prompt. This means your characters look consistent across different illustrations — Theron always has his red cloak, Elara always carries her silver staff.
Creating Variations¶
Hidden Feature
Click Generate Image again to create a different scene or variation. Each generation is independent — you can build a gallery of multiple moments from the same session.
Managing Images¶
- Images are displayed newest-first in the gallery
- Delete individual images or clear all images for a session
- The most recent image is used as the hero image in PDF exports
Requirements¶
- Apple Silicon Mac (M1+) — image generation runs directly on your Mac, no external service needed
- A connected AI assistant is needed to craft the scene description (see Settings)
- Image quality can be tuned via Steps and Guidance Scale in Settings
Next: Export Your Work →