Parameters and Configuration

Parameters control how content is generated. Each model has unique parameters you can adjust.

Parameter Types

Input (Text)

Simple text field:

{
  type: "input",
  label: "Prompt",
  placeholder: "Describe what you want to generate",
  required: true
}

Textarea (Long Text)

For long prompts or detailed descriptions:

{
  type: "textarea",
  label: "Detailed Description",
  rows: 5,
  maxLength: 1000
}

Select (Selection)

List of predefined options:

{
  type: "select",
  label: "Resolution",
  options: [
    { value: "512", label: "512x512" },
    { value: "1024", label: "1024x1024" }
  ]
}

Slider

Numeric value with range:

{
  type: "slider",
  label: "Guidance Scale",
  min: 1,
  max: 20,
  step: 0.5,
  default: 7.5
}

Checkbox

Boolean value:

{
  type: "checkbox",
  label: "High Quality",
  default: false
}

File

File upload (images, audio, etc.):

{
  type: "file",
  label: "Reference Image",
  accept: "image/*",
  maxSize: "10MB"
}

Common Parameters

For Image Generation

Width and Height

Image size in pixels:

• 512x512: Fast, good for testing
• 1024x1024: Quality/speed balance
• 2048x2048: High quality, slower

Guidance Scale

How closely to follow the prompt:

• 1-5: Creative, less faithful to prompt
• 7-10: Recommended balance
• 15-20: Very faithful to prompt, less creative

Steps

Number of iterations:

• 20-30: Fast, basic quality
• 50: Recommended balance
• 100+: High quality, very slow

Seed

Number to reproduce results:

• Random: Different each time
• Fixed: Same result with same seed
• Negative: Generates variations

For Video Generation

Duration

Video duration in seconds:

• 2-5 seconds: Fast, good quality
• 5-10 seconds: Recommended balance
• 10+ seconds: More processing time

FPS

Frames per second:

• 24 fps: Cinematic standard
• 30 fps: Smooth and natural
• 60 fps: Very smooth, heavier

Motion

Movement intensity:

• Low: Subtle movement
• Medium: Natural movement
• High: Dynamic movement

Advanced Parameters

Negative Prompt

Describe what to avoid in generation:

Example:
"blurry, low quality, distorted, watermark"

Sampler

Sampling algorithm:

• Euler: Fast and efficient
• DDIM: More controlled
• DPM++: High quality

CFG Scale

Classifier-free guidance:

• Controls prompt strength
• Higher values = more faithful to prompt
• Lower values = more creative

Parameter Validation

The system automatically validates:

• ✅ Required fields: Must be completed
• ✅ Ranges: Values within limits
• ✅ Types: Correct format
• ✅ Dependencies: Related parameters

Error Messages

If a parameter is invalid:

• Error message is shown
• Field is marked in red
• Explains what's wrong
• Suggests how to fix it

Parameter Presets

Using Presets

Presets are saved configurations:

1. Select a preset from the selector
2. Parameters load automatically
3. You can adjust afterwards if needed

Creating Presets

1. Configure parameters as desired
2. Click "Save as preset"
3. Assign a descriptive name
4. Preset becomes available for future generations

Default Presets

Each model includes recommended presets:

• Fast: Optimized for speed
• Quality: Optimized for maximum quality
• Balance: Balance between speed and quality

Parameters by Image Field

Some models allow multiple images:

Image (Main Image)

• Main reference image
• Used as base for generation
• Required for some models

Reference Images

• Additional reference images
• Help guide the style
• Optional in most cases

Control Images

• Control images for structure
• Guide composition
• Specific to some models

Grid Configuration

You can customize how results are displayed:

Columns

• 2 columns: Expanded view
• 3 columns: Recommended balance
• 4+ columns: Compact view

Size

• Compact: More items visible
• Normal: Standard size
• Large: Detailed view

Best Practices

Start with Default Values

✅ Recommended:

• Use default values first
• Experiment with one parameter at a time
• Save configurations that work

Understand Parameters

• Read each parameter's description
• Consult model examples
• Experiment with different values
• Compare results

Optimize for Your Use Case

• Speed: Lower values
• Quality: Higher values
• Creativity: Medium-low guidance scale
• Fidelity: High guidance scale

Troubleshooting

Parameters don't apply

• Verify model supports the parameter
• Check value is in valid range
• Reload page if necessary

Unexpected results

• Review all parameters
• Try with default values
• Consult model documentation

Next Steps

• Prompt templates
• Content library
• Token system