chatgpt image prompts
ChatGPT Image image-to-video prompt
You already have a strong still from ChatGPT Image and need controlled motion without warping the subject.
Example prompt
Keyframe: product bottle on marble, moody rim light. Motion: slow 15° orbit, condensation drip, background bokeh static, 5s, no text, no morphing label.
Copy into ChatGPT Image / GPT Image; adjust product names and hex colors.
Steps
Step 1
Export max-resolution still
Use PNG; avoid heavy compression before the video model ingests the frame.
Step 2
Split still vs motion prompts
Do not restyle in the motion step—only describe camera and micro-movement.
Step 3
Stitch in editor
Add sound and supers in post; keep generative motion subtle to reduce artifacts.
Multi-step pipeline: Workflow index · Recipe library
Related prompts
FAQ
- Why separate still and motion prompts?
- Video models reinterpret style if you merge prompts; keep look locked in the still.