Prompt Box

Prompt Box Overview

The Prompt Box is the main area where you describe what you want to see in your generated image. It also gives you access to various settings that let you control the visual style, subject consistency, aspect ratio, colors, and more.

Prompt Box in its default state.

Elements of the Prompt Box

1. Prompt Field – Where you type your text prompt — a description of what you want the AI to generate.

2. Image Upload – Lets you upload an image to use in these different situations:

   1. As a Character or Style reference, or to Describe the content of the uploaded image.

   2. To edit the image in various ways such as Magic Fill, Remix, Upscale, Remove or Replace the background, or edit it in Canvas.

3. Model / Render Speed / Number of Images – Lets you select which AI model/version to use (e.g., Ideogram 3.0 vs earlier versions), the quality/rendering speed, and how many output images are generated per prompt.

The Model + variations menu.

1. Aspect Ratio – Select the format of your image (e.g. square, portrait or landscape.)

The Aspect ratio menu.

1. Style – Choose a visual style or a Style Reference to guide the overall look. You can pick from global mode (like Realistic, Design, General), from a list of specific built-in presets or use/create a custom one.

The Style popup window.

1. Character Reference – Use a reference image to guide how a person or character looks across multiple generations.

2. Color Palette – Choose a color palette for the generated image. Guides overall color tones.

The Color palette menu.

1. More Options – Click the More icon to access more settings:

   1. Tile – Generate seamless tiles (Model 1.0 only)

   2. Seed Number – Reproduce the same image output that was done before.

   3. Negative Prompt – Tell the AI what to avoid.

The More menu.

1. Magic Prompt – A tool/setting that optimizes and embellishes the original prompt (when needed), to maximize the variety and add more details to the images generated.

The Magic Prompt menu.

1. Private Generation – A visibility toggle that makes your image visible only to you or to the public. Requires a Plus plan or higher to generate images privately.

The Visibility menu.

1. Generate – Button that starts the generation process based on your prompt and selected settings.

---

Prompt Box in Different Modes

When you select a Style or a Character Reference, or Remix an image, a small thumbnail appears inside the Prompt Box. These thumbnails visually represent the current selections applied to your prompt.

Prompt Box with option thumbnails such as Remix, Style, or Character references.

Collapsed Prompt Box showing option thumbnails grouped on the right side.

1. Style Reference thumbnail – Shows that a custom style reference image is selected and active.

2. Character Reference thumbnail – Indicates that a character image reference is selected.

3. Remix thumbnail – Shows the image being remixed and an image weight (e.g., 50).

4. Grouped thumbnails – In some views, the Prompt Box collapses and shows only the prompt and thumbnails of your selected style, character, or remix image.

Other Possible Thumbnails

• Upscale thumbnail – Shows that a custom style reference image is selected and active.

You can click on the Character or Remix thumbnail to bring options in a dropdown menu, or on the Style thumbnail to bring up the Style popup window.

---

Importing an Image

When you import an image—either by clicking the Image Upload icon , pasting an image directly into the prompt field, or dragging it onto the page—a popup window appears with several options for how to use or edit the image.

The available options when importing an image.

Available actions

• Character – Use the uploaded image to create a Character Reference, preserving the subject’s identity and facial features across multiple generations.

• Style – Use the image as a Style Reference, capturing its aesthetic, color scheme, and visual tone to influence future generations.

• Describe – Automatically generate a detailed text prompt that describes the uploaded image. Useful for re-creating similar content.

• Magic Fill – Edit the image by removing or replacing areas intelligently. The AI fills the masked region based on your text prompt.

• Remix – Reimagine the uploaded image while preserving its composition. Adjust the Remix weight to balance between originality and stylistic influence.

• Upscale – Enhance the resolution and details of the image for sharper, higher-quality results.

• Open in Canvas – Launch the image in Canvas for manual editing, masking, and combining multiple assets.

• Replace BG – Automatically detect and remove the background, replacing it with a new scene or color based on your prompt. You can click on the right end dropdown arrow to select Remove BG which removes the background and make it transparent (PNG).

Tip: Choosing any of these options will transfer you to the corresponding feature page or editor view automatically, depending on the selected action.

---