The Workflow
8 steps · click a step number to mark it done
Open Sora2 and have it ready before you start
Collect or create reference images that illustrate the characters, props, and environment for your scene and save them to your computer
Open ChatGPT, paste the Sora2 documentation link, and ask it to write a 12‑second cinematic prompt that includes camera moves, lighting, mood, and references to your uploaded images
Copy the detailed prompt that ChatGPT generates
In the Art List (or Sora2) interface, select the Sora2 Pro model, set the duration to 12 seconds, choose your desired aspect ratio, paste the prompt, and click Generate
Review the multiple video variations Sora2 returns and pick the one that most closely matches your vision
If the result needs tweaks, tell ChatGPT specific changes (e.g., increase camera orbit speed, add motion blur) and ask it to update the prompt
Paste the revised prompt back into Sora2 and regenerate the video
Repeat the refinement loop until the video looks cinematic, then export the final video file