Unstable Diffusion User Guide

Welcome to Unstable Diffusion, where we bring your creative visions to life with the power of artificial intelligence. Whether you're a new user or an advanced AI artist, this guide is designed to help you navigate our platform and make the most out of our extensive features.

1. Getting Started

1.1 Credits System

Our service operates on a two-tiered credit system.

• Slow Credits: These refresh daily and allow you to generate images at a Relaxed speed.
• Fast Credits: These can be obtained with a paid subscription and allow for Turbo-speed generations.

The number of credits required for a generation corresponds to the number of images produced. For instance, 1 image requires 1 credit, and 4 images require 4 credits.

New users receive 12 Fast credits upon sign-up. To access more Fast credits, purchase a monthly subscription.

2. Basic Features

2.1 Prompt Input

Enter your creative concept into the Prompt input field. This helps guide the AI in generating your desired image.

2.2 Model Selection

Select your preferred model from the Model dropdown. There are currently 4 models, each specialized on their own domain but with large overlaps with each other:

• Merlin: A general model, capable of many art styles and topics.
• Echo: Focused on photography, but also capable of other styles like illustration.
• Izanagi: Anime and manga inspired model.
• Pan: A model focused specifically on creating anthropomorphic characters and scenes, in an anime aesthetic.

Trying your prompts and experimenting with all 4 models will allow you to create art is many diverse and varying styles!

2.3 Aspect Ratio

Determine the layout of your art by sliding the aspect ratio slider among three options: landscape, square, and portrait.

2.4 Number of Images

You can generate up to 4 images at a time by selecting the appropriate button under the Number of Images section.

2.5 Generation Speed

Toggle the Generation Speed switch to select between Relaxed and Turbo mode, the latter roughly halves the time taken to generate an image.

2.6 Create Button

When ready, click the Create button. You'll see your image gradually forming as indicated by the circular progress bar.

2.7 Exclude Input

If there are recurring elements that you do not want to appear in your image, you can mention them in the Exclude input field.

3. Advanced Features

Navigate to the Settings tab at the top of the generation menu for additional options to fine-tune your image.

3.1 Samplers

The sampler influences the aesthetics of your image. Options include:

• Diverse Composition: Amplifies the color variety.
• Dynamic Contrast: Enhances the contrast between light and dark elements.
• Complementary Composition: Generate images featuring colors complementary to a chosen lighting color.
• Color Guidance: Create images predominantly featuring a chosen color.
• Dual Color Guidance: Predominantly features a chosen lighting color, and attempts to exclude a selected negative color.

3.2 Sliders

• Sampler Strength: Adjusts the impact of the chosen sampler.
• High Frequency Detail: Introduces more fine detail.
• Saturation: Alters the color vibrancy and saturation.

3.3 Reset Settings Button

To return to default settings, click the Reset Settings button.

4. Usage Guidelines and Consequences

• If a prompt potentially violates our guidelines, the generation will be denied.
• If generated images violate our guidelines, they will not be served.

5. Additional Notes

Right-click to download images at full resolution or to permanently delete them. Be aware, once deleted, images are permanently lost.

Next to each generated image, you will find the Regenerate and Reuse Settings buttons. These options allow you to reapply the generation details or regenerate images using the prior settings.

Your settings will persist even when you navigate away or refresh the page.

Pro and Premium tier subscribers are conferred commercial rights of the images they generate.

We hope this guide provides a clear understanding of our service. Enjoy your journey in creating AI-powered art with Unstable Diffusion.