Image Generation¶

AIA supports AI-powered image generation through various models, enabling you to create images from text descriptions, modify existing images, and integrate visual content generation into your workflows.

Supported Models¶

DALL-E Models (OpenAI)¶

DALL-E 3: Latest and most capable image generation model
DALL-E 2: Previous generation, still available and capable

Image Model Capabilities¶

# Check available image generation models
aia --available_models text_to_image

# Example output:
# - dall-e-3 (openai) text to image
# - dall-e-2 (openai) text to image

Basic Image Generation¶

Simple Image Generation¶

# Generate an image with default settings
aia --model dall-e-3 "A serene mountain lake at sunset"

# Generate with specific size
aia --model dall-e-3 --image_size 1024x1024 "Modern office workspace"

# Generate with quality settings
aia --model dall-e-3 --image_quality hd "Professional headshot"

Image Configuration Options¶

Image Size (`--image_size`, `--is`)¶

# Square formats
aia --image_size 1024x1024 "Square image prompt"
aia --is 512x512 "Smaller square image"

# Landscape formats  
aia --image_size 1792x1024 "Wide landscape image"
aia --is 1344x768 "Medium landscape"

# Portrait formats
aia --image_size 1024x1792 "Tall portrait image"
aia --is 768x1344 "Medium portrait"

Available sizes: - Square: 256x256, 512x512, 1024x1024 - Landscape: 1792x1024, 1344x768 - Portrait: 1024x1792, 768x1344

Image Quality (`--image_quality`, `--iq`)¶

# Standard quality (faster, less expensive)
aia --image_quality standard "Quick concept image"

# HD quality (better detail, more expensive)  
aia --image_quality hd "High-quality marketing image"
aia --iq hd "Detailed technical diagram"

Quality options: - standard: Good quality, faster generation, lower cost - hd: Enhanced detail and resolution, slower, higher cost

Image Style (`--style`, `--image_style`)¶

# Vivid style (hyper-real, dramatic colors)
aia --image_style vivid "Dramatic sunset over city skyline"

# Natural style (more natural, less stylized)
aia --image_style natural "Realistic portrait of a person reading"
aia --style natural "Documentary-style photograph"

Style options: - vivid: Hyper-real and dramatic images - natural: More natural, less stylized results

Advanced Image Generation¶

Using Prompts for Image Generation¶

Create reusable image generation prompts:

# ~/.prompts/product_photography.txt
//config model dall-e-3
//config image_size 1024x1024
//config image_quality hd
//config image_style natural

# Product Photography Generator

Generate a professional product photograph of: <%= product %>

Style requirements:
- Clean, minimalist background
- Professional lighting
- Commercial photography style
- <%= lighting || "Soft, even lighting" %>
- <%= background || "White background" %>

Additional specifications:
- Angle: <%= angle || "45-degree angle" %>
- Context: <%= context || "Isolated product shot" %>
- Mood: <%= mood || "Clean and professional" %>

# Use the prompt
aia product_photography --product "wireless headphones" --lighting "dramatic side lighting"

Complex Image Descriptions¶

# ~/.prompts/detailed_scene.txt
//config model dall-e-3
//config image_size 1792x1024
//config image_quality hd
//config image_style vivid

# Detailed Scene Generator

Create a detailed image of: <%= scene_type %>

## Visual Elements:
- Setting: <%= setting %>
- Time of day: <%= time_of_day || "golden hour" %>
- Weather: <%= weather || "clear" %>
- Color palette: <%= colors || "warm and inviting" %>

## Composition:
- Perspective: <%= perspective || "eye level" %>
- Focal point: <%= focal_point %>
- Depth of field: <%= depth || "shallow depth of field" %>

## Style and Mood:
- Art style: <%= art_style || "photorealistic" %>
- Mood: <%= mood || "peaceful and serene" %>
- Technical quality: <%= quality || "professional photography" %>

Generate a <%= scene_type %> scene with <%= focal_point %> as the main subject, 
set in <%= setting %> during <%= time_of_day %>.

Image Series Generation¶

# ~/.prompts/image_series.txt
//config model dall-e-3
//config image_size 1024x1024

# Image Series Generator

//ruby
series_theme = '<%= theme %>'
variations = ['<%= var1 %>', '<%= var2 %>', '<%= var3 %>']
base_prompt = '<%= base_description %>'

puts "Generating #{variations.length} variations of #{series_theme}:"
puts

variations.each_with_index do |variation, index|
  puts "## Image #{index + 1}: #{variation.capitalize}"
  puts "#{base_prompt} featuring #{variation}."
  puts "Style: Consistent with series theme of #{series_theme}"
  puts
end

# Generate a series
aia image_series \
  --theme "modern architecture" \
  --base_description "Professional architectural photograph" \
  --var1 "glass and steel skyscraper" \
  --var2 "minimalist residential house" \
  --var3 "contemporary office building"