AI Image Widget

The AI Image widget lets you create and edit images using AI. Type a description of what you want, pick a model, and get an image in seconds. You can also upload an existing image to edit it, remove its background, or upscale it to a higher resolution.

#How to Use

#Generate a New Image

1. Make sure the mode is set to Generate (the default).
2. Type a description of the image you want in the prompt field - for example, "a cozy cabin in the snow."
3. Choose an AI model from the model selector (each model has different strengths and costs).
4. Pick an aspect ratio if you need a specific shape (landscape, portrait, or square).
5. Click the Run button to generate your image.

#Edit an Existing Image

1. Switch to Edit mode.
2. Drag and drop an image into the widget, or click the drop zone to browse your files.
3. Describe the changes you want in the prompt field - for example, "add a sunset in the background."
4. Click Run to apply the edits.

If you drop an image while in Generate mode, the widget automatically switches to Edit mode.

#Remove a Background

1. Switch to Remove BG mode.
2. Upload the image you want to process.
3. Click Run - no prompt is needed for this mode.

The widget will return your image with the background removed.

#Upscale an Image

1. Switch to Upscale mode.
2. Upload the image you want to enlarge.
3. Pick an upscale factor - 2x, 3x, or 4x - using the factor button in the toolbar. The default is 2x.
4. Click Run - no prompt is needed for this mode.

The widget returns a higher-resolution version of your image at the chosen factor.

#Models

The widget offers several AI models, each with different strengths:

• FLUX.2 Flash - fast and affordable, great for quick ideas (default)
• FLUX.2 - realistic textures and lighting
• FLUX.2 Pro - high resolution with precise colors
• Nano Banana - photorealistic with accurate text rendering
• Nano Banana 2 - enhanced version with more detail
• Nano Banana Pro - best quality, lighting, and details
• Seedream 4.5 - 4K resolution with great typography
• GPT Image 1.5 - complex scenes, signs, and infographics
• Grok Imagine - portraits, logos, and multiple styles
• Qwen Image - multi-line text in English and Chinese
• BiRefNet - dedicated background removal model

Each model shows its credit cost before you run it, so you always know what a generation will cost.

#Aspect Ratios

When generating a new image, you can cycle through five aspect ratios:

• 1:1 - square (default)
• 16:9 - wide landscape
• 9:16 - tall portrait
• 4:3 - standard landscape
• 3:4 - standard portrait

#Saving Your Image

After generating or editing an image, click the download button to save it to your device.