Image-to-video prompts are not for describing the picture — the model already sees it. Describe **what happens next**. Stick to one explicit subject motion plus one camera move. Multi-action prompts cause warping and jitter.

In-depth articles in this hub

Each card below is a full breakdown — click in to read the complete article:

Best for

  • Animating an AI still for an intro
  • Turning product images into GIFs
  • Adding 3-second motion to a portrait

Quick reference: example prompts

Quick sample prompts you can copy and adapt (full walkthroughs are in the articles above):

Subject smiles slightly

The subject in the image slowly smiles and looks back at camera, slight head tilt, very subtle camera push-in, no other motion, 4 seconds

Wind in the scene

A gentle wind blows across the scene, leaves and hair move slightly, slow static camera, soft natural motion, 5 seconds

Product rotation

The product in the image slowly rotates 30 degrees on its central axis, soft studio lighting unchanged, no other element moves, 4 seconds