How to generate images with AI

PixelFiddler lets you generate brand-new images from text descriptions using AI. Instead of searching stock photo libraries or hiring a designer for every background and decorative element, you can describe what you need in plain language and get a ready-to-use image in seconds.

This guide covers the two ways to generate AI images, how to write prompts that get great results, and practical use cases for your e-commerce store.

What AI image generation does

You type a text description — called a prompt — and PixelFiddler’s AI model creates an original image that matches your description. The result is a unique image you can use in your product listings, banners, social media posts, and templates.

For example, a prompt like “a clean white marble surface with soft natural lighting, minimalist product photography style” produces a polished background you can place behind your product photos.

Use real photos for your actual products

AI-generated images work best for backgrounds, decorative elements, and banner artwork. For your main product images, always use real photographs. Shoppers expect to see exactly what they are buying, and authentic product photos build trust and reduce returns.

Where you can generate AI images

There are two places in PixelFiddler where you can create AI images: the Media Library and Templates. Each serves a different purpose.

---

Option 1: Generate in the Media Library

Use this when you want to create a standalone image you can download, reuse across listings, or add to templates later.

How to generate

1. Open the Media Library for your storage
2. Click the AI Generate button in the header toolbar
3. Enter your text prompt in the prompt field (up to 1,000 characters)
4. Choose your desired image size from the dropdown
5. Select the format — PNG or JPG
6. Review the token cost displayed below the settings
7. Click Generate

Available image sizes

Pick the size that best fits how you plan to use the image:

Size	Best for
512 x 512	Thumbnails, small social media graphics
1024 x 576	Wide banners, landscape headers
576 x 1024	Tall banners, portrait-oriented displays
1024 x 768	Standard landscape backgrounds
768 x 1024	Standard portrait backgrounds
1024 x 1024	Square social media posts, general-purpose backgrounds
2048 x 2048	High-resolution backgrounds, large print-ready graphics

Note

Larger images use more tokens. The exact token cost is always shown before you confirm generation, so you can decide whether the size is worth the cost for your use case.

Format options

• PNG — Lossless quality, larger file size. Best when you need crisp edges or plan to layer the image over other content.
• JPG — Smaller file size with slight compression. Best for photographic-style backgrounds and web use.

After generation

Once the AI finishes creating your image, it is automatically saved to your media library with a timestamp in the filename. From there you can:

• Use it as a background in product composites
• Download it for use outside PixelFiddler
• Add it to templates as an image layer
• Apply further transformations like resizing, cropping, or effects

---

Option 2: Generate inside a Template

Use this when you want AI to create an image dynamically every time a template is rendered. This is ideal for backgrounds or decorative elements that should vary or update automatically.

How to set it up

1. Open or create a Template in the template editor
2. Add a new layer and choose the AI Image layer type
3. Set the layer dimensions (width and height)
4. Enter a default prompt that describes the image you want
5. Choose the AI model tier — Basic or Quality
6. Position the layer on the canvas and adjust opacity or blend mode as needed

When the template is rendered, PixelFiddler generates a fresh image from the prompt and composites it into the final output along with your other layers.

Note

The prompt set on the AI Image layer is a default. If your template uses variables, you can override the prompt at render time to generate different images from the same template.

Good use cases for AI layers in templates

• Dynamic backgrounds — Generate varied backdrop textures or scenes behind your product images
• Decorative elements — Create borders, patterns, or artistic accents for promotional banners
• Placeholder visuals — Fill template slots with representative images while you finalize your designs

---

Writing effective prompts

The quality of your generated image depends largely on how you write your prompt. Here are practical guidelines for getting the best results.

Be specific and descriptive

Vague prompts produce vague images. The more detail you provide, the closer the result will match your vision.

Instead of this	Try this
”background"	"a clean white marble surface with soft natural lighting"
"nature scene"	"a sunlit forest floor covered in green moss with warm golden light filtering through trees"
"pattern"	"a seamless geometric pattern in navy blue and gold with art deco style”

Include style direction

Tell the AI what visual style you are going for. This helps it match the look and feel of your brand.

• “minimalist product photography style”
• “flat illustration with pastel colors”
• “watercolor texture with soft edges”
• “luxury brand aesthetic with dark tones”

Mention lighting, colors, and textures

These three elements have the biggest impact on whether the generated image will work well alongside your product photos.

• Lighting — “soft diffused lighting”, “bright studio lighting”, “warm sunset glow”
• Colors — “neutral earth tones”, “vibrant coral and teal”, “monochrome gray palette”
• Textures — “smooth matte finish”, “rough linen fabric texture”, “glossy reflective surface”

Keep it focused

A prompt that tries to describe too many things at once often produces cluttered results. Focus on one clear scene or concept per image.

Start simple, then refine

Begin with a short prompt, review the result, and then add more detail in your next attempt. This iterative approach helps you learn what descriptions work best for the kind of images your store needs.

---

Use cases for e-commerce

Here are the most common ways online sellers use AI image generation in PixelFiddler.

Lifestyle backgrounds for product compositing

Generate realistic surfaces and environments to place your product photos on. A jewelry seller might generate “a smooth dark slate surface with soft directional lighting from the left” and then composite their product on top using a template.

Banner and promotional artwork

Create eye-catching backgrounds for seasonal sales, holiday promotions, or social media campaigns without hiring a designer for every banner.

Template backgrounds

Build reusable templates with AI-generated background layers. When you need a new variation, just update the prompt and render again.

Placeholder images during design

While building out your templates and store layout, generate placeholder images that approximate your final vision. Replace them with real assets when ready.

---

Tracking generation progress

After you click Generate, PixelFiddler shows a status indicator so you can follow the progress. Generation typically takes a few seconds depending on the image size and current demand. You can continue working in other parts of the app while the image is being created — it will appear in your media library once complete.

Caution

AI image generation consumes tokens from your account balance. Always check the token cost displayed in the generation dialog before confirming. Larger image sizes and the Quality model tier cost more tokens per generation.

---

Next steps

Now that you know how to generate images with AI, explore these related features:

Edit images with AI Use AI to modify existing images -- remove backgrounds, change colors, and more

Templates overview Learn how to build reusable image templates with multiple layers

Models and pricing Compare AI model tiers and understand token costs