Goal

• Given the text of a storyboard for a presentation, generate an image prompt for each slide and save the result to demo_images.md

Workflow

Generate Image Prompts

• For each slide, generate an image prompt in the following format, adding a description in <DESCRIPTION>
• The prompt should be like:

  CHARACTERS
  <DESCRIPTION>
  
  STYLE ANCHOR:
  Clean, modern vector-illustration style, flat colors, soft gradients, consistent
  lighting, no harsh shadows, slightly rounded shapes, minimal detail, professional
  corporate look, designed for a slide deck, white background, high contrast, 16:9
  aspect ratio, no text, no watermark.
  
  SETTINGS:
  - No photorealistic elements or textures
  - No slide titles, headers, or explanatory text blocks
  - No decorative elements that don't serve the concept
  - No complex scenes with many small details
  - Minimize the number of distinct elements (5-7 maximum per image)
  
  CAMERA: Medium shot, straight-on.
  

Create the Output File

• Write the generated prompts to demo_images.md so it can be used with the command:

  > ./helpers_root/dev_scripts_helpers/documentation/generate_images.py \
     -i demo_images.md \
     --dst_dir demo_images.md.figs
  

Verification

• Confirm demo_images.md has one prompt per storyboard slide
• Confirm each prompt follows the CHARACTERS / STYLE ANCHOR / SETTINGS / CAMERA format
• Run generate_images.py against demo_images.md and confirm it succeeds