AI Image Generator Pipeline

Overview

The AI Image Generator Pipeline is a production-ready automation skill that transforms webhook triggers into fully optimized, published images across your digital ecosystem. This skill orchestrates a complete workflow: receive webhook events → intelligently generate or enhance prompts → call DALL-E 3 or Flux AI → optimize images for web/mobile → upload to WordPress, AWS S3, or Cloudinary → notify stakeholders.

Why this matters: Content teams waste 40% of time on image sourcing and optimization. This skill eliminates manual steps, ensures consistency, and scales visual content production without hiring designers.

Integrations: WordPress (REST API), AWS S3, Cloudinary, Slack (notifications), Zapier (webhook triggers), Make.com, Google Drive (backup storage), Stripe (usage billing).

Quick Start

Try these example prompts immediately to see the skill in action:

Generate a hero image for a blog post about sustainable fashion with 
DALL-E 3, optimize for web (1920x1080), and upload to my WordPress 
media library with alt text "sustainable fashion hero image".

Webhook trigger: When a new Shopify product is created, generate 
a product lifestyle image using Flux AI with the product description 
as the prompt, compress to 500KB, and upload to S3 with CDN URL.

Batch generate 5 variations of a minimalist coffee shop interior 
using DALL-E, apply Instagram-style filters, add watermarks, and 
create a WordPress gallery post with all variations.

Monitor incoming webhook events from my email marketing platform. 
For each campaign, generate a hero banner image from the campaign 
subject line, optimize for mobile (800x600), and auto-insert into 
the email template's image placeholder.

Capabilities

1. Intelligent Prompt Generation

• Accepts raw webhook data and transforms it into detailed, style-consistent prompts
• Maintains prompt templates for brand consistency (e.g., "minimalist tech startup aesthetic")
• Supports dynamic variables: {product_name}, {brand_color}, {target_audience}
• Automatic prompt enhancement with style descriptors, lighting, composition cues

Example:

Input webhook: {"product": "Blue Wireless Earbuds", "category": "electronics"}
Generated prompt: "Premium blue wireless earbuds on white marble surface, 
studio lighting, minimalist composition, product photography, 8k, trending on behance"

2. Multi-Model Image Generation

• DALL-E 3: Superior quality, better prompt adherence, ideal for branded content
• Flux AI: Faster generation, excellent for batch operations, lower cost
• Model selection logic: automatic based on quality/speed requirements or user preference
• Supports size variants: 1024x1024, 1024x1792, 1792x1024 (DALL-E); flexible (Flux)
• Retry logic with exponential backoff for API failures

3. Advanced Image Optimization

• Format conversion: JPEG (web), PNG (transparency), WebP (modern browsers), AVIF (next-gen)
• Compression: Intelligent quality settings (0-100), target file sizes (50KB-5MB)
• Resizing: Responsive breakpoints (mobile: 375px, tablet: 768px, desktop: 1920px)
• Watermarking: Brand logo overlay, text watermarks, opacity control
• Filters & effects: Saturation boost, contrast enhancement, color grading presets
• Metadata embedding: SEO-friendly alt text, EXIF data, copyright notices

Example:

Optimize for WordPress: 
- Generate 3 sizes: 300x300 (thumbnail), 800x600 (medium), 1920x1080 (full)
- Convert to WebP with quality 85
- Compress to max 200KB per image
- Add brand watermark (bottom-right, 20% opacity)
- Embed alt text and SEO keywords

4. Multi-Destination Publishing

• WordPress: Direct REST API upload to media library, auto-attach to posts, set featured images
• AWS S3: Bucket organization by date/category, CloudFront CDN integration, access control
• Cloudinary: Cloud-based optimization, dynamic URL transformations, version control
• Google Drive: Automated backup, folder organization, sharing permissions
• Custom webhooks: POST optimized image URLs to your own systems

5. Webhook Event Processing

• Listens on custom webhook endpoint (auto-generated)
• Supports: Shopify (product creation), WordPress (post publish), Stripe (payment events), custom JSON
• Webhook signature verification (HMAC-SHA256) for security
• Automatic retry on failed deliveries (3 attempts, exponential backoff)
• Webhook history and audit logs

6. Batch Operations & Scheduling

• Process multiple images in parallel (configurable concurrency: 1-10)
• Schedule generation at off-peak hours to save API costs
• Cron-based triggers: "Generate product images every morning at 2 AM"
• Dry-run mode to preview outputs before publishing

Configuration

Required Environment Variables

# Image Generation APIs
OPENAI_API_KEY=sk-...              # DALL-E 3 access
FLUX_API_KEY=flux-...              # Flux AI access (optional)

# WordPress Integration
WORDPRESS_URL=https://mysite.com   # Your WordPress domain
WORDPRESS_API_TOKEN=abc123xyz      # Application password (WordPress 5.6+)
WORDPRESS_USER=automation          # API user account

# AWS S3 Integration
AWS_S3_BUCKET=my-images-bucket     # S3 bucket name
AWS_ACCESS_KEY_ID=AKIA...          # IAM access key
AWS_SECRET_ACCESS_KEY=...          # IAM secret key
AWS_REGION=us-east-1               # S3 region

# Cloudinary (Optional)
CLOUDINARY_CLOUD_NAME=mycloud      # Cloudinary account
CLOUDINARY_API_KEY=...             # Cloudinary API key
CLOUDINARY_API_SECRET=...          # Cloudinary API secret

# Slack Notifications (Optional)
SLACK_WEBHOOK_URL=https://hooks... # Slack incoming webhook
SLACK_CHANNEL=#content-alerts      # Notification channel

# Security
WEBHOOK_SECRET=your-secret-key     # HMAC signing key for webhooks

Configuration Options

# config.yaml example
image_generation:
  model: "dalle3"                   # "dalle3" or "flux"
  quality: "high"                   # "standard", "high"
  size: "1024x1024"                 # DALL-E sizes
  style: "vivid"                    # "vivid" or "natural"

optimization:
  formats: ["webp", "jpg"]          # Output formats
  quality: 85                       # JPEG/WebP quality (0-100)
  max_file_size_kb: 200             # Max compressed size
  add_watermark: true               # Embed brand watermark
  watermark_path: "./assets/logo.png"
  watermark_opacity: 0.2

publishing:
  primary_destination: "s3"         # "wordpress", "s3", "cloudinary"
  secondary_destinations: ["wordpress"]
  s3_folder_structure: "images/{date}/{category}"
  wordpress_post_status: "draft"    # Auto-publish or save as draft
  generate_srcset: true             # Create responsive image sets

webhook:
  port: 3000
  path: "/webhook/images"
  timeout_ms: 30000
  max_retries: 3

notifications:
  slack_enabled: true
  notify_on: ["success", "failure"]
  include_image_preview: true

Setup Instructions

1. Install dependencies:

   npm install openai @aws-sdk/client-s3 cloudinary axios
   # or: pip install openai boto3 cloudinary requests
   

2. Set environment variables:

   export OPENAI_API_KEY="sk-..."
   export WORDPRESS_URL="https://mysite.com"
   # ... (see above for all vars)
   

3. Generate webhook endpoint:

   # The skill auto-generates a unique webhook URL:
   # https://openclaw.ai/webhooks/ai-image-gen-{skill_id}
   # Copy this and add it to your Shopify/WordPress/Stripe webhook settings
   

4. Test the connection:

   curl -X POST https://openclaw.ai/webhooks/ai-image-gen-{skill_id} \
     -H "Content-Type: application/json" \
     -d '{"product": "Test Product", "description": "A blue widget"}'
   

Example Outputs

Scenario 1: E-Commerce Product Image Generation

Input (Shopify webhook):

{
  "event": "product.created",
  "product": {
    "id": 12345,
    "title": "Wireless Noise-Canceling Headphones",
    "description": "Premium audio with 30-hour battery",
    "vendor": "AudioTech Pro"
  }
}

Generated Prompt:

"Premium wireless noise-canceling headphones in matte black and silver, 
studio product photography, white background, professional lighting, 
high-end audio equipment aesthetic, 8k quality, trending on product hunt"

Output:

WordPress Media Library:
├── audiotech-pro-headphones-1024x1024.webp (150 KB)
├── audiotech-pro-headphones-800x600.webp (98 KB)
├── audiotech-pro-headphones-300x300.webp (32 KB)
└── Meta: alt="Premium wireless noise-canceling headphones by AudioTech Pro"

AWS S3:
└── s3://my-images/images/2024-01-15/electronics/
    ├── audiotech-pro-headphones-full.webp
    └── audiotech-pro-headphones-thumb.webp

Slack Notification:
✅ Image generated successfully
Product: Wireless Noise-Canceling Headphones
Model: DALL-E 3
Size: 1024x1024
File size: 150 KB
WordPress: https://mysite.com/wp-content/uploads/2024/01/audiotech-pro.webp
CDN URL: https://d123.cloudfront.net/images/2024-01-15/audiotech-pro.webp

Scenario 2: Blog Content Image Generation

Input (WordPress post publish event):

{
  "event": "post.published",
  "post": {
    "id": 42,
    "title": "10 Sustainable Fashion Trends for 2024",
    "excerpt": "Eco-friendly clothing is no longer niche. Here are the trends defining sustainable fashion.",
    "category": "Fashion"
  }
}

Generated Output:

Image 1 (Hero banner - 1920x1080):
- Prompt: "Sustainable fashion 2024 trends, eco-friendly clothing, 
  organic materials, minimalist aesthetic, diverse models, natural lighting, 
  editorial photography, trending on vogue"
- Format: WebP, 85% quality, 180 KB
- Alt text: "Sustainable fashion trends 2024 - eco-friendly clothing collection"
- Destination: WordPress featured image + S3 backup

Image 2 (Social media square - 1080x1080):
- Same prompt, cropped for Instagram
- Format: WebP, 90% quality, 120 KB
- Watermark: "SustainableFashionBlog.com" (bottom-right, 15% opacity)

Batch Output:
├── featured-image-hero-1920x1080.webp
├── featured-image-hero-1080x1080.webp (Instagram)
├── featured-image-hero-800x600.webp (mobile)
└── featured-image-hero-300x300.webp (thumbnail)

Tips & Best Practices

1. Prompt Engineering for Consistency

• Create brand style guides as prompt templates:

  Template: "{product_name} in {brand_color}, {style} aesthetic, 
  {lighting}, {composition}, {quality_level}, trending on {platform}"
  

• Use consistent descriptors: "studio lighting" vs "natural light", "minimalist" vs "maximalist"
• Test prompts manually first; refine before automating

2. Cost Optimization

• DALL-E 3: $0.04-0.10 per image (1024x1024). Best for hero/featured images.
• Flux: $0.01-0.03 per image. Best for batch operations and thumbnails.
• Strategy: Use Flux for thumbnails/social, DALL-E 3 for hero images
• Set up cost alerts: notify when monthly spend exceeds $100
• Use dry-run mode to preview before publishing (saves API calls)

3. Image Optimization Best Practices

• WebP format: 25-35% smaller than JPEG, supported by 95% of browsers
• Responsive images: Generate 3 sizes minimum (mobile/tablet/desktop)
• File size targets: Thumbnails <50KB, medium <150KB, full <300KB
• Watermarking: Use 15-25% opacity to avoid obscuring content
• SEO: Always include descriptive alt text (80-120 characters)

4. Webhook Reliability

• Implement idempotency: use product ID + timestamp as unique key to prevent duplicates
• Monitor webhook logs: check failed deliveries in the skill dashboard
• Set up Slack alerts for generation failures
• Test webhook signatures in development before production

5. WordPress Integration Tips

• Use "draft" status initially; review before auto-publishing
• Set featured images automatically: _thumbnail_id meta field
• Organize media by folder: /2024/01/products/ structure
• Leverage WordPress image sizes: register custom sizes in theme
• Use SEO plugins (Yoast) to auto-fill meta descriptions from alt text

6. Batch Processing Efficiency

• Process in parallel (concurrency: 5-10) to reduce total time
• Schedule batch jobs at 2-4 AM to use cheaper off-peak API rates
• Monitor API rate limits: DALL-E (3 req/min free tier, 50 req/min paid)
• Use caching: store generated prompts to avoid re-generating

7. Brand Consistency Across Channels

• Create separate prompt templates for each platform:
  • Instagram: square (1080x1080), vibrant colors, trendy aesthetic
  • LinkedIn: professional (1200x627), corporate look, minimal text
  • Website: responsive (1920x1080), aligned with brand guidelines
• Use consistent color palettes: reference brand hex codes in prompts
• A/B test: generate 2 variations, track engagement to refine prompts

Safety & Guardrails

What This Skill Will NOT Do

1. Copyright & IP Protection:

   • Will NOT generate images that mimic copyrighted artwork or trademarked designs
   • Will NOT bypass DALL-E's content policy (no celebrity likenesses without permission)
   • Will NOT create images of real people without explicit consent
   • Recommendation: Review generated images for IP issues before publishing

2. Harmful Content:

   • Will NOT generate violent, explicit, or hateful content
   • Will NOT create deepfakes or misleading images
   • Will NOT generate images that violate OpenAI/Flux terms of service
   • Built-in content filters block prohibited prompts

3. **Unauthorized Publishing