AI Image Generation

The first step in creating your custom image is selecting the Image Style. Soulkyn offers a variety of styles to match your artistic preferences. Here's a rundown of the available styles:

• Anime Enhanced: This style provides sharper, more vivid details with rich colors. Perfect for anime fans who want a polished, vibrant look that pops.
• Anime Original: For a classic, hand-drawn feel, this style offers a softer and more traditional anime aesthetic, capturing that nostalgic anime vibe.
• Realistic Original: If lifelike images are what you're after, this style delivers clean, realistic representation with crisp details, making characters look like they've stepped out of reality.
• Realistic Cinématographique: This style gives your images a cinematic, film-like appearance with dramatic lighting and artistic flair. It's ideal for creating intense or atmospheric scenes that feel like a movie still.

How to Choose: From the Image Generation interface, click the Image Style dropdown and select your preferred style to set the vibe for your custom creation.

On Soulkyn, you can generate images by selecting Empty (where you have complete control) or by using various Premade Prompts tailored for specific image types.

Empty (User Manages Everything):
If you love having full control over every element—like clothing, background, and prompts—select Empty. This option allows you to manually tweak every detail, perfect for when you have a specific vision in mind and want to manage all aspects of the image creation process.

Premade Prompts Options:
For a quicker setup, Soulkyn offers a range of premade prompts to match your needs. These pre-configured options help you get started faster, with settings for Full Body, Portrait, and NSFW images that cater to different artistic intentions.

• Full Body: Choose this option if you want a wider shot with the camera pulled back, showing the entire body of your character—ideal for showcasing outfits, action poses, or full-body compositions.
• Portrait: This preset is perfect for closer shots that focus on the character's face and upper body, making it great for profile pictures or when you want to capture expressions in detail.
• NSFW Premade Prompts: Quickly set the scene for mature or intimate images with these prompts, which position your character in more sensual or provocative settings, tailored for NSFW content.

How to Select a Premade Prompt:
After setting the number of images, browse through the premade prompt options and select the one that fits your vision best. It's a simple way to get the look you want without having to manually configure each element.

After choosing a Premade Prompt (or going with 'Empty'), you can take personalization to the next level by applying User Presets for Background, Clothes, and Additional Prompts. These presets let you fine-tune specific elements of the image, ensuring it perfectly matches your vision.

Background Presets: Set the scene by choosing a preset background that matches the mood you're aiming for—whether it's a serene beach, a dark dungeon, or a bustling cityscape.

Clothes Presets: Customize your character's outfit to fit their personality or the scenario you're creating. From casual wear to fantasy armor, these presets help you dress your persona for the occasion.

Additional Prompts: Add extra details or instructions to refine the image further. These prompts can guide the AI to include specific props, adjust the character's pose, or tweak the lighting for a perfect touch.

To learn more about setting up your own presets, check out our preset guide.

Sometimes certain keywords in the physical traits can cause our text-to-image model to "undress" your character. For example, using a tag like "pierced nipples" will likely remove clothing to display that feature.

How to Fix It:
A simple solution is to remove that specific trait from the physical attributes and instead include it in the default clothing preset for your character (clothes have less weight). You can also create multiple clothing presets—one with these features and one without—to switch between as needed.

This way, you can maintain control over the appearance while keeping things SFW when intended.

Grouping: Use parentheses () to group words or phrases.

Modifiers:

• + or ++ to emphasize.
• - or -- to de-emphasize.
• More than two + or - are reduced to ++ or --.

Note: Ensure all parentheses are properly closed.

Example Prompt:

(a cute cat)++, red fur, small eyes, cute+, (playing with red ball)-

Note: This will only work in Additonal Prompts and not in Physical Traits

Understanding the difference between images in chat and image generation can help you get the results you're looking for. While both features create visuals, they serve distinct purposes, responding to different inputs and contexts.

Image in Chat:
Images in chat are designed to illustrate the ongoing conversation. They don't just respond to your specific prompt; they consider the full context of the chat, including the current conversation logs, the AI's last response, and the character's set location and clothes. This means you don't always need to use "show me" or "send me" prompts—simply clicking "picturize" will generate a contextual image based on the current scene, even if there's no graphic description in your message. They aren't designed for exact image requests. Instead, they give a visual representation of the chat flow, making it feel more dynamic and immersive.

Image Generation:
For a precise image of your character, the image generation tool is your go-to. Unlike images in chat, image generation doesn't rely on contextual details and won't be influenced by ongoing conversations. It's ideal for creating specific poses or scenes, like "showing her butt while bending over a table," without pulling from the chat context.

Quick Guide:
If your goal is to generate a specific image of a character, use the image generation tool. If your goal is to create a contextual illustration of the current chat, use the image in chat feature.

If your character's images keep generating with a second character, it's likely due to conflicting physical or clothing tags in your character setup. This can happen when the AI model interprets the tags as describing multiple individuals instead of one.

Common Causes:
Conflicting tags like "short hair" and "long hair" or "red hair" and "black hair" can make the model think it's meant to create two separate characters. Similarly, contradictory clothing descriptions, such as "large breasts" and "small breasts," may lead to this issue.

How to Fix It:
Here are some ways to resolve the problem:

• Add Solo Tags: Include prompts like solo female focus or solo male focus in the physical or clothing attributes. These tags emphasize that the image should feature a single character.
• Avoid Contradictions: Remove conflicting physical or clothing tags from your character's setup. Instead of using contradictory traits, opt for combinations that make sense together, such as black hair with red highlights or short hair with bangs.

By ensuring that your tags are consistent and specific, you can guide the AI to generate images that focus solely on your character without unintentionally adding another.