Image Nodes

Image nodes are where your visual creativity comes to life. They use AI models to generate images from text prompts or transform existing images.

What Are Image Nodes?

An Image node with the Nano Banana model selected for AI image generation.

Image nodes generate images using AI. They can:

• 🎨 Generate from text — Describe what you want, get an image
• 🖼️ Transform images — Use reference images to guide generation
• 📐 Control output — Set size, style, and quality

Adding an Image Node

1. Open the node sidebar on the left
2. Find Image in the list
3. Drag it onto your canvas

Image Node Inputs

Image nodes accept these inputs:

Connecting a Prompt

1. Add a Text Node with your prompt
2. Drag from the Text Node's output to the Image Node's prompt input
3. Or type directly in the Image Node's prompt field

Connecting a Reference Image

1. Add an Upload Node with your reference image
2. Drag from the Upload Node's output to the Image Node's image input
3. The AI will use this as a reference

Image Node Settings

Model Selection

Choose from available AI models:

The model selector showing all available image AI models.

Aspect Ratio

Control the output dimensions:

Additional Settings

Text-to-Image Workflow

The most common workflow:

[Text Node] → [Image Node]
Your prompt → Generated image

Example

1. Add a Text Node
2. Write: "A cozy coffee shop interior, morning light, warm tones, photography"
3. Add an Image Node
4. Connect Text → Image
5. Select model (e.g., Flux 2 Dev)
6. Click Run

Image-to-Image Workflow

Use a reference image to guide generation:

[Upload Node] → [Image Node]
Reference image → Transformed image

Example: Style Transfer

1. Add an Upload Node
2. Upload your reference image
3. Add an Image Node
4. Connect Upload → Image (to the image input)
5. Add a Text Node with style description
6. Connect Text → Image (to the prompt input)
7. Run to generate styled version

Example: Image Editing with Flux Kontext

1. Upload your image to edit
2. Use Flux Kontext Pro or Flux Kontext Max
3. Prompt describes the change: "Change the background to a beach sunset"
4. Run to get edited image

Model-Specific Features

Flux Models

• Excellent for photorealistic images
• Great text rendering
• Professional quality output
• Kontext variants for editing

Nano Banana

• Best for text in images
• Consistent multi-image generation
• Logical compositions
• Pro version for highest quality

SDXL

• Wide range of artistic styles
• Good for illustrations
• Many style presets
• Community model support

Understanding Settings

Guidance Scale

Controls how literally the AI follows your prompt:

Recommendation: Start with 7-10 for most uses.

Steps

More steps = higher quality but slower:

Recommendation: 30-40 steps for most uses.

Seed

The random seed controls reproducibility:

• Random (default) — Different image each time
• Fixed number — Same image with same prompt/settings

Use fixed seeds when you want to:

• Reproduce a specific result
• Make small prompt adjustments
• Create variations

Viewing Results

When generation completes:

1. Preview appears in the node
2. Click the preview to see full size
3. Download button to save locally
4. Image is auto-saved to your Gallery

Chaining Image Nodes

Connect image outputs to other nodes:

Image → Tool

[Image Node] → [Upscaler Tool]
Generated image → Higher resolution version

Image → Video

[Image Node] → [Video Node]
Generated image → Animated video

Image → Image

[Image Node 1] → [Image Node 2 (Kontext)]
Initial image → Edited version

Tips for Better Images

Prompt Tips

• Be specific about subject, setting, lighting
• Include style keywords
• Mention camera/photography terms for realism
• See Prompt Engineering

Model Selection

• Testing ideas → Qwen Image (cheap, fast)
• Artistic work → SDXL
• Professional output → Flux 2 Pro
• Text in images → Nano Banana
• Editing images → Flux Kontext

Quality Tips

• Use higher steps for final output
• Match aspect ratio to your use case
• Use reference images for consistency
• Iterate with cheap models, finalize with quality models

Common Issues

Image Doesn't Match Prompt

• Increase guidance scale
• Be more specific in prompt
• Try a different model
• Check for conflicting instructions

Image Quality Low

• Increase steps
• Use a higher-quality model
• Check aspect ratio (extreme ratios may reduce quality)

Text in Image Wrong

• Use Nano Banana or Nano Banana Pro
• Put text in quotes in your prompt
• Keep text short and simple
• Be explicit about text placement

Example Prompts by Use Case

Product Photography

Professional product photography of a luxury watch, 
white background, soft studio lighting, 
commercial quality, high detail

Social Media Content

Lifestyle photo of coffee and laptop on wooden desk, 
morning sunlight through window, cozy atmosphere, 
Instagram aesthetic, warm tones

Marketing Visual

Modern tech startup office, diverse team collaborating, 
bright and professional, corporate photography style,
authentic and engaging

Next Steps

• Video Nodes — Animate your images
• Tool Nodes — Enhance your images
• Flux Models Prompting — Master Flux prompts