Wildcards let you create dynamic prompts with random variations. Instead of typing the same prompt multiple times with different subjects, use wildcards for automatic randomization.
{category:option1|option2|option3}
Example:
A {animal:cat|dog|bird} playing {instrument:piano|guitar|drums} in a {location:park|studio|street}
Generates one of:
- "A cat playing piano in a park"
- "A dog playing guitar in a studio"
- "A bird playing drums in a street"
- (Any combination of the options)
{category}
Example:
A {person} using a {tool} to {action} the {object}
Uses built-in categories:
- {animal}, {person}, {profession}, {instrument}, {location}
- {weather}, {time}, {emotion}, {action}, {color}
- {vehicle}, {tool}, {object}
| Category | Options |
|---|---|
| {animal} | cat, dog, bird, horse, rabbit, fox, deer, wolf |
| {person} | man, woman, child, elder, teenager, artist, scientist |
| {profession} | detective, chef, doctor, teacher, engineer, musician, artist |
| {instrument} | piano, guitar, violin, drums, saxophone, flute |
| {location} | park, studio, street, forest, beach, city, room |
| {weather} | sunny, rainy, snowy, foggy, stormy, cloudy |
| {time} | morning, noon, afternoon, evening, night, midnight |
| {emotion} | happy, sad, angry, surprised, calm, excited, pensive |
| {action} | walking, running, dancing, working, playing, creating |
| {color} | red, blue, green, yellow, purple, orange, black, white |
| {vehicle} | car, motorcycle, bicycle, truck, bus, train, boat |
| {tool} | hammer, wrench, paintbrush, camera, telescope, microscope |
| {object} | book, ball, box, bottle, phone, computer, chair |
Input: A {profession:detective|chef|scientist} in their {location:office|kitchen|laboratory} at {time:dawn|midnight|noon}
Possible outputs:
- "A detective in their office at dawn"
- "A chef in their kitchen at midnight"
- "A scientist in their laboratory at noon"
Input: A {color:red|blue|black} {vehicle:sports car|motorcycle|truck} racing through a {weather:rainy|snowy|foggy} {location:city|mountain pass|desert}
Possible outputs:
- "A red sports car racing through a rainy city"
- "A blue motorcycle racing through a snowy mountain pass"
- "A black truck racing through a foggy desert"
Input: A {person} {emotion:happily|sadly|angrily} holding a {color} {object}
Uses predefined {person} and {object} with inline emotions and colors.
Input: {profession:astronaut|pilot|diver} explores {location:alien planet|underwater cave|arctic tundra} during {time:sunrise|sunset|night} with {tool:scanner|camera|equipment}
Creates varied exploration scenes automatically.
- Wildcard Detection: Node scans your prompt for
{...}syntax - Random Selection: Picks one option from each wildcard
- Replacement: Substitutes wildcards with chosen options
- Expansion: Sends processed prompt to LLM for detail expansion
The node shows what was chosen:
Original Input:
A {animal:cat|dog|bird} playing {instrument:piano|guitar}
Wildcard Replacements:
- animal: cat
- instrument: piano
Processed Prompt:
A cat playing piano
When generating multiple variations, wildcards are re-rolled for each:
Variation 1: A cat playing piano in a park
Variation 2: A dog playing guitar in a studio
Variation 3: A bird playing drums in a street
Each variation gets different random selections!
✅ Good: {location:park|beach|forest} (all outdoor)
❌ Bad: {location:park|spaceship|microscope} (unrelated)
✅ Good: A {animal:cat|dog|bird} (all singular)
❌ Bad: A {animal:cat|dogs|birds} (mixed singular/plural)
Use inline wildcards when you need specific options not in predefined lists:
{character:wizard|warrior|rogue|healer}
Input: A {profession} in a {location}
Preset: noir
Tier: advanced
Result: Noir-style expansion of the random profession/location combo
Input: A {vehicle} racing through {weather} conditions
Shot Size: wide shot
Camera Movement: tracking shot
Preset: action
Result: Action-style wide tracking shot of random vehicle/weather
Wildcards work GREAT with random preset:
Input: A {animal} {action}
Preset: random
Tier: enhanced
Result:
- Random animal chosen (e.g., "fox")
- Random action chosen (e.g., "dancing")
- Random cinematography applied
- Output: "A fox dancing" with random aesthetic choices
- Check syntax:
{name:opt1|opt2}(colon and pipes) - No spaces in curly braces:
{animal}not{ animal }
- Not a bug! Each run picks randomly
- Use variations (num_variations: 3) to get multiple
If you see {animal} in the output:
- Typo in category name
- Category not in predefined list
- Use inline format:
{animal:cat|dog|bird}
You can combine wildcards with concepts:
Input: {profession} using {tool} to fix a {object} in a {weather} {time}
Creates complex scenarios like:
"engineer using wrench to fix a computer in a rainy morning"
Input: A {profession} in dramatic pose
Preset: cinematic
Positive Keywords: myLoRA_trigger
Tier: advanced
Result: Random profession, cinematic style, with LoRA trigger
Input: A {vehicle} in motion
Preset: action
Shot Size: wide shot
Camera Movement: tracking shot
Time of Day: sunset time
Lens: wide-angle lens
Result: Random vehicle, but specific cinematography
Wildcards enable:
- ✅ Dynamic prompt variations
- ✅ Quick batch generation with variety
- ✅ A/B testing different concepts
- ✅ Creative exploration
- ✅ Efficient prompt iteration
Syntax:
- Inline:
{name:opt1|opt2|opt3} - Predefined:
{category}
Works with:
- All presets (especially good with "random")
- All tiers
- All modes (text-to-video, image-to-video)
- Both Standard and Advanced nodes
- Multiple variations
Try it!
Input: A {animal} {action} in a {location} during {time}
Generate 3 variations and see the variety!