- Add Pillow dependency for image processing
- Remove text instructions from fal.ai prompts (generate clean images)
- Add semi-transparent dark background box behind text for readability
- Overlay full title text with white text and black outline
- Add proper line spacing between text lines
- Fix FAL_KEY environment variable setup
- Add image URL logging to console output during batch processing
- Remove unused h2-prompts file
- Implement ImageGenerator class with hero and content image generation
- Add image theme prompt generation and caching
- Integrate with fal.ai flux-1/schnell API
- Add image upload to storage (Bunny CDN)
- Add image injection into HTML content
- Add test script for image generation
- Update database models and repositories for image fields
- Fix API usage: use arguments parameter and image_size object format