Media¶

Portraits, videos, visuals

Media

Commands:
  /img <name|location|me>    Generate portrait/location image
  /vid <name|me> [prompt]    Generate video from portrait
  /visual <name> key=val     Set visual attributes
  /tag <name> <tags>         Set character tags
  /scene [character]         Generate scene composite
  /viewer                    Open portrait viewer panel
  /viewrefresh               Force refresh viewer state

Media — Portraits¶

Character and location images

Media — Portraits

/img <name>          Generate character portrait
/img me              Generate player portrait
/img location        Generate location background
/img all             Batch portraits for all NPCs here
/keyroll <name>      Reroll face seed (new look)
/keyroll me          Reroll player face seed
/noimages            Toggle auto-portrait generation

Images are generated via ComfyUI (--comfy flag).
Portraits are cached — regenerate only when
visual state changes. Use /visual to modify
appearance, /tag for descriptive tags.

/noimages disables auto-gen on room entry; manual
/img still works. Toggle again to re-enable.

Media — Visuals¶

Appearance and tags

Media — Visuals

/visual <name> key=value     Set visual attributes
/visual <name> --clear       Clear all visuals
/tag <name> tag1, tag2       Set character tags
/tag <name> --clear tag1     Clear then set tags

Visual keys: hair_color, hair_style, eye_color,
build, skin_tone, clothing, expression, etc.
Changes trigger portrait regeneration.

Media — Video¶

Video generation

Media — Video

/vid <name|me> [prompt]      Generate video
/vid <name> --draft          Low-res quick preview
/vid <name> --fast           Medium quality

Generates a short animated clip from a character
portrait using WAN 2.1 I2V via ComfyUI.
Requires ComfyUI with WAN model loaded.

Media — Scenes¶

Scene composites and narrative snapshots

Media — Scenes

/scene [character]   Generate scene composite image
/snapshot (/snap)    Cinematic narrative snapshot

/scene combines player + NPC portraits into a single
scene image. Requires --rapid-aio flag.
Uses the focused character if none specified.
Auto-triggers during adult activities.

/snapshot asks the narrator to paint the current moment
cinematically — 3-5 sentences of atmosphere, setting,
and character presence. Does NOT advance the story;
it's a pure observational beat. Not saved to history.