This video was generated using the Elements feature of Kling AI.
2 subjects + 1 environment = a new scene
It gets the action right, but details slip through.
Here’s the breakdown:
↳ How it works:
#1 → open kling
#2 → go to image-to-video
#3 → select elements
#4 → upload images + text prompt
#5 → generate → iterate → refine
↳ Kling prompt structure
[shot type + direction], [subjects + details + action], [environment], [mood/vibe], [end scene details], [style]
↳ Prompt example:
"A cinematic tracking shot follows a drag race between a black ferrari f40 and a white mclaren on a winding mountain pass. the camera moves with the cars through tight turns, tires skidding and spraying water as they accelerate. high-contrast lighting, ending with a wide shot of a vast valley clearing. shot in cinematic 4k."
This interesting example was created by Rory Flynn. My go-to for anything related to AI image and video generation.