This runs the AI vision analysis only, then builds the 3 JSON prompts. It does not generate images.
Sign in before running the AI photo analysis.