Configure your videoChoose the cinematic style, camera movements and mood
2
Click - it's automatic!The prompt is generated, copied and the AI tool opens. Paste with Ctrl+V.
3
Generate your videoPaste the prompt in Kling, Runway, Sora or your preferred tool
4
Direct the gaze explicitlyThe model predicts visually consistent pixels, not narratively consistent actions: it does not know who should look at whom, or why. Use ultra-directive prompts about the gaze: "character turns head LEFT, eyes locked on [object/person], chin pointing toward [target]" - the more you describe the physical mechanics of the gaze, the less the model improvises.
🎉 Optimized video prompt
1 Start from your reference image
→2 Describe the motion
→3 Copy into Kling
Your image already sets the style, the light and the colours. The prompt only describes what moves. One step follows the other · within a step, everything is simultaneous.