Image Generation

Image Generation has arrived at NovelAI!

What is Image Generation?

The NovelAI Diffusion Anime & Furry image generation experience is unique and tailored to give you a creative tool to illustrate your visions without limitations, allowing you to paint the stories of your imagination.

Use tags to define the visual characteristics of your character or composition (or you can let AI interpret your words if you prefer!)

Generate images on NovelAI with our own custom NovelAI Diffusion Models, based on Stable Diffusion:

NovelAI Diffusion Anime (Curated) Good baseline quality and predictable subject matter.

NovelAI Diffusion Anime (Full) The expanded training set allows for a wider variety of generations.

NovelAI Diffusion Furry (Beta) Beta model to easily create Furry and Anthropomorphic Animal themed content. This model is not a complete work, and will be expanded upon!

You can access Image Generation straight from the Dashboard or from the User Menu accessible from the goose icon on the Library Sidebar.

Where to access Image Generation.

How it works

Stable Diffusion works by generating an image based on your text prompt, starting with only noise, and gradually enhancing the image until there is no noise left at all.

How the AI creates an image by using diffusion

The AI connects your text to images, generating a new composition every single time you prompt it.

The NovelAI Diffusion image generation experience is unique, tailored to give you a creative tool to illustrate your visions without limitations, allowing you to paint the stories of your imagination. Easily define characteristics of your character and composition with tags and get a little closer to that vision in your head (or not, but now it’s completely up to your choice!)

The AI will suggest tags that it knows!

Settings at a glance

All NovelAI Image Generation Settings at a glance (Subject to change!)

Image Resolution

You can generate images at a variety of aspect ratios and sizes. Use the dropdown to select the option best suited to your vision!

Goose Tip: Press the X between the resolution numbers to quicky swap height and width.

Number of Images Number of Images

Below the resolution setting, you can select how many images you want generated at once. The maximum amount of images possible depends on your chosen resolution; The higher the resolution, the fewer images you will be able to generate at once. For example, Normal-sized images can be generated up to 6 at a time, while Large-sized images can be generated in batches of 4 at most.

Goose Tip: Increasing the Number of Images also increases the Anlas cost of each generation. Even if you're an Opus subscriber, generating more than one image at once will always cost Anlas.

Normal Generations

Steps define the number of iterations the AI should refine for.

Prompt Guidance indicates how strongly the AI interprets your prompt. At high Guidance, the AI follows your prompt closely, with finer detail and sharpness. Lower Guidance gives the AI more creative freedom.

Collapsable Settings

Goose Tip: The AI Settings box can be collapsed to free some space for other UI elements. While in that state, you can still change the Steps and Guidance values by clicking and typing their values!

Image2Image Generation

Strength influences how much the AI will change the image composition. Lower Strength values will result in finished images looking closer to the original.

Noise may allow the AI to generate more details, but can cause strange artifacts if too high.

Note: Strength & Noise only apply when an image is used to generate with, so these settings do not appear in the menu unless that condition is met!

Topics to browse