Image to video AI: make any picture move

Why start from an image?

A text prompt gives the model freedom, and freedom means each result looks different. An image removes that uncertainty. If you need a specific product, a character you designed or an exact composition, this is the mode to use.

How to do it

1. Upload a clear, well-lit image, in the shape you want the video to be.
2. Write about what changes, not what is already there. “She turns to the window and smiles” beats a long description of the woman.
3. Say what the camera does: a slow push in, or no movement at all.
4. Run a short test, then render the longer version.

What makes a good source image

• Sharp and evenly lit. Blurry images give blurry motion.
• Room to move. Leave space around the subject so the camera and action have somewhere to go.
• A simple subject. Crowds and fine detail are harder to animate well.
• Your own image. Use pictures you have the right to use, and get consent for anything showing a real person.

Going further with two images

Supply a first and a last frame and the model creates the movement between them. It works well for before-and-after shots, product reveals and transitions that land exactly where you want. If you want the same subject in entirely new scenes, use reference images instead, explained in our guide to consistent characters.

Common uses

Product shots for ads, character art for stories, portraits and landscapes that gently come alive, and AI-generated images turned into motion. For personal pictures, see photo to video AI.