AI Editor

AI Editor in Effect House allows you to train and customize your own AI model to stylize your effects. You can select from available AI models and use prompt to describe the style you want to apply to your image.

Key Features

• Preset Library – A variety of predefined styles for easy selection and preview
• Image Generator – You can input descriptive text (prompts) to customize and generate an AI model that you can apply to your effect
• Preview Media – Upload your own images to preview your model

Access AI Editor

To open the AI Editor window:

1. Open Effect House and navigate to the Tool Bar
2. Click AI Editor to open the AI Editor window

3. Alternatively, click Add Object in the Hierarchy panel, go to 2D, and select AI Image

4. You can also access the AI Editor from an AI Texture. Go to the Inspector panel and under Art Style, click the Base Model field to reopen AI Editor.

You can now start using AI Editor to create your own AI model.

AI Editor Key Features

• Image Generator – A chat panel with built-in customization settings where you can input descriptive text (prompts) to generate an AI model
• Preset Library – A variety of predefined models that you can use as is or customize using the image generator
• Preview Media – A section where you can upload your own images or choose from a library of images to preview your model

Image Generator

The image generator outputs an AI model based on the base model, text prompt, reference image, and style selections. It enables you to generate a unique base model that can be used as AI Texture or AI Image within effects.

Generating your AI Model

Base Model

To generate your AI model, start by selecting a base model. Base models serve as the foundation for generating stylized effects. Each base model is a pre-trained AI framework with distinct characteristics, trained on various datasets to produce specific styles. These models define the core structure, color schemes, and rendering techniques before any additional stylization is applied. Learn more about the different base models.

Text Prompt

The chat panel allows you to enter your text prompt or the style of image you want to create from the base model. For example, "A futuristic city skyline at sunset with flying cars." Use clear and specific language to describe your vision.

note

Read How to Write a Well-Crafted Prompt for inspiration and additional guidance, and make sure your prompt adheres to our Community Guidelines.

Reference Image

To supplement your text prompt, you can upload an image to use as a style reference for the AI model.

Customization Settings (Optional)

To further customize your model, you can adjust the parameters within More Settings. These parameters can change depending on the base model you choose.

1. Prompt Strength: Refers to the degree of influence that your text prompt has on the final AI-generated image. It determines how strictly the AI follows your input prompt versus how much creative freedom it takes to generate variations.
2. Depth: Adjusts the perception of depth and dimensionality in the generated image
3. Edge Sensitivity: Determines how strongly the AI detects and preserves edges and contours within the image
4. Subject Similarity: Controls how closely the AI-generated image resembles the original input (for example, reference image or base content)
5. Face Similarity: Adjusts how much the AI prioritizes preserving facial features from the original input
6. Cutout: Enables or disables the background behind the subject

Generate

Generates the final results based on the text prompt, reference image, and optional customization settings.

Edit, Regenerate, and Download

• Edit: Re-edits the prompt to it's original prompt text
• Regenerate: Regenerates the model based on the prompt
• Download: Allows you to download all images as an asset

Preset Library

There are two main tabs in the Preset Library. Trending displays models created by the Effect House community. Base Models display the same base models that are in the Preset Library. You can use any of the models in these tabs as is and apply it to your effect, or customize it with the image generator.

Preview Media Library

You can click Preview Media to select or upload the media you want to use to preview your AI model.

1. Click Preview Media

2. Select the media you want to use for preview or upload your own media
3. Click Done. You are now taken back to the Image Generator.

4. Click Generate. Your AI model is now applied to your newly selected media. In this example, you can see the images are generated based on your preview image.

5. From here, you can click on an image and compare it with the original media, apply the model to your effect, or download the image as an asset.