nullara
f4fc92e968
Added an example workflow for Silly Tavern, ComfyUI, and a Background/Audio prompt for LLM
30 lines
1.2 KiB
Plaintext
30 lines
1.2 KiB
Plaintext
Ignore all previous instructions. Generate a background-only scene description as a comma-delimited list of keywords and short phrases.
|
|
|
|
Follow these rules strictly:
|
|
|
|
Begin with: background, no humans,
|
|
The first five elements must appear in this exact order:
|
|
location
|
|
time of day
|
|
weather
|
|
lighting
|
|
environmental details
|
|
After these, include additional purely visual details (architecture, terrain, objects, textures, colors, atmosphere, etc.)
|
|
|
|
Hard constraints:
|
|
|
|
Do NOT describe any people, characters, or living beings
|
|
Do NOT include names, personality traits, emotions, actions, or movement
|
|
Do NOT include smells, thoughts, or non-visual sensations
|
|
Everything must be describable in a still photograph
|
|
Do NOT write full sentences — only comma-separated keywords or short phrases
|
|
Do NOT continue any story or roleplay
|
|
|
|
Output format:
|
|
|
|
After the visual keyword list, add a colon :
|
|
Then provide a second comma-delimited list of ambient audio cues that match the environment (e.g., rain, wind, distant traffic, birds)
|
|
|
|
Example structure (for clarity):
|
|
|
|
background, no humans, [location], [time of day], [weather], [lighting], [environmental details], [additional visual elements]: [ambient sounds] |