Home / Settings Guide

Settings Guide

Master all generation parameters to create the perfect AI artwork

The /settings Command

The easiest way to configure all your generation settings is through the /settings command. It opens an interactive UI where you can adjust everything in one place:

What you can configure:
  • Model selection
  • Image size
  • Generation steps
  • CFG scale
  • Seed
  • Sampler
  • Default prompt & negative prompt
  • Hi-res settings
  • LoRA selection and weights

Tip: Use /settings for easy access to all options. Below you'll find detailed explanations of each parameter.

📊 Generation Steps

What are Steps?

Steps determine how many iterations the AI performs to refine your image. More steps generally mean higher quality but take longer to generate.

Recommended Values:
  • 10-15 steps: Quick previews, lower quality
  • 20 steps: Default, good balance of speed and quality
  • 25-30 steps: Higher quality, maximum recommended

Note: Values above 30 rarely improve results and can cause over-processing. Stick to 20-30 for best results.

🎯 CFG Scale

What is CFG Scale?

CFG (Classifier-Free Guidance) scale controls how closely the AI follows your prompt. Higher values mean stricter adherence to the prompt, while lower values allow more creative freedom.

Recommended Values:
  • 1-2: Very creative, may ignore parts of prompt
  • 3-5: Balanced creativity and prompt adherence (recommended)
  • 6-7: Strong prompt adherence, good for specific requests
  • 8-10: Very strict, may cause over-saturation or artifacts

Tip: Most users get the best results with CFG 3-5. Higher values can cause oversaturation.

🔄 Samplers

What are Samplers?

Samplers are algorithms that control how the AI generates images from noise. Different samplers produce different results and have varying speeds.

Available Samplers:
  • Euler a: Fast, good quality, most popular choice (default)
  • Euler: Non-ancestral variant, more predictable results
  • DPM++ 2M Karras: Excellent quality/speed balance, great for detailed work
  • DPM++ SDE Karras: Good for artistic/detailed outputs, slightly slower
  • DDIM: Deterministic, excellent for testing and reproducibility

Tip: Select samplers via the /settings menu for easy access to all options.

📐 Image Size

Choosing the Right Size

The AI models are optimized for specific resolutions. Using the recommended sizes ensures the best quality output.

Optimized Sizes:
  • 1024x1024: Square, default, best overall quality
  • 1152x896: Landscape orientation
  • 896x1152: Portrait orientation
  • 1216x832: Wide landscape
  • 832x1216: Tall portrait
  • 1344x768: Ultrawide landscape
  • 768x1344: Ultra tall portrait

Note: Using non-standard sizes may result in distortions or quality loss. Select your preferred size in the /settings menu.

🎲 Seed

What is a Seed?

The seed is a number that determines the initial random noise pattern. Using the same seed with identical settings produces the same image, allowing you to reproduce results.

How to use:
  • -1: Random seed each time (default behavior)
  • Any positive number: Specific seed for reproducible results
  • Seeds are shown in the generation info of each image
  • Use the same seed to create variations by changing other parameters

Tip: Configure the seed in the /settings menu, or set it to -1 for random seeds.

📝 Default Prompts

Default Prompt

Set a prefix that will be automatically added to all your generations. Useful for quality tags or consistent style.

Examples:
  • masterpiece, best quality, highly detailed
  • by artist_name, digital art
  • furry, anthro, detailed fur

Tip: Set your default prompts via the /settings menu.

Negative Prompt

Specify what you DON'T want in your images. The AI will try to avoid these elements.

Common negative prompts:
  • low quality, blurry, jpeg artifacts
  • bad anatomy, deformed, extra limbs
  • watermark, signature, text
  • simple background (if you want detailed backgrounds)

Note: The bot has a built-in safety negative prompt that's always applied. Set your custom negative prompt via the /settings menu.

🔍 Hi-Res Settings

What is Hi-Res Mode?

Hi-res mode generates your image at a lower resolution first, then upscales and refines it. This produces more detailed images but takes longer.

Hi-Res Parameters:
  • Upscaler: Algorithm used for upscaling (Latent, R-ESRGAN, etc.)
  • Scale: How much to upscale (1.5x, 2x, etc.)
  • Denoising: How much to refine during upscale (0.3-0.7 recommended)
  • Steps: Number of refinement steps in the second pass

Tip: Enable and configure Hi-Res mode in the /settings menu.

💧 Watermark

Custom Watermarks

Add a personalized watermark to your generations to protect your artwork or brand it.

Watermark Options:
  • Text: Your custom watermark text
  • Position: Where to place it (corners, center, etc.)
  • Opacity: How visible the watermark is (0-100%)
  • Font Size: Size of the watermark text
  • Color: Watermark color

Tip: Configure watermarks via the /settings menu.

🖼️ Img2Img Settings

Image-to-Image Transformation

Configure how the AI transforms reference images when using /img2img.

Img2Img Parameters:
  • Denoising Strength (0.0-1.0):
    • 0.1-0.3: Minor changes, keeps most of original
    • 0.4-0.6: Moderate transformation
    • 0.7-1.0: Major changes, less like original
  • Resize Mode: How to handle size mismatches
  • Scale: Output size multiplier

Tip: Configure Img2Img settings via the /settings menu.

💡 Pro Tips

🎯

Start Simple

Begin with default settings and adjust one parameter at a time to understand its effect.

🔄

Use Seeds

Found a good result? Save the seed and experiment with prompt variations while keeping the composition.

⚡

Quick Previews

Use 15-20 steps for quick previews, then increase for final renders.

📏

Match Model to Size

Always use the optimized sizes for the best results. Non-standard sizes often cause distortions.

🎨

Combine LoRAs

Try combining a character LoRA with a style LoRA for unique results.

📊

CFG Sweet Spot

For most generations, CFG 3-5 gives the best balance of quality and creativity.