The Brief
Exploring how JSON-based prompting delivers structured, cinematic outputs with far greater precision and control than traditional text prompts.
CREATIVE APPROACH
While standard prompts can leave room for interpretation, JSON prompting allowed me to break down each scene into modular elements — camera, subject, motion, emotion, and environment. This structured approach created consistent results across multi-scene narratives while maintaining flexibility for creative iteration.
I broke down prompt components and tested each core prompt variable one at a time: Subject, Style, Camera/Lens, Lighting, and Composition.
WORKFLOW
Veo 3 – Core video generation with JSON-structured prompts
JSON Structuring – Modular breakdown of dynamic actions
ChatGPT – Prompt iteration and refinement for operator-grade clarity
Premiere Pro – Sequencing and assembling JSON-generated clips
Leveraged optimized prompts to generate visually consistent frames for motion graphics previz workflows, style frame development, storyboards, and pitch visuals.
PROMPT
I structured the JSON prompt by breaking the animation into discrete elements—speaker motion, camera moves, lighting, and environment—assigning precise parameters for timing, rotation, and reflections. This modular approach enabled synchronized, cinematic visuals and rapid iteration, ensuring control and consistency throughout the sequence.
Main Goal
A demonstration of how JSON prompting transforms Veo 3 into a precise storytelling tool — enabling cinematic flow, operator-level control, and scalable workflows that go beyond what’s possible with traditional text prompts.
KEY TAKEAWAYS
I played a key role in helping to build The AI Artist Loft by assisting their team to generate its foundational prompt library and generating iterative image banks to showcase visual range. I organized workflow systems for prompt testing and optimization, using a variety of AI tools to refine output quality and consistency.
