Why start from a photo instead of a prompt
When you already have the exact product, person or place you need on screen, a photo is a far more reliable starting point than words. The maker locks the opening frame to your image, so the thing in the video is unmistakably your thing — same packaging, same face, same room.
That matters for brand work. A text prompt can get you something that looks like your product; a photo gets you your product.
Which photos animate best
Sharp, well-lit images with a clear subject and some depth do best. Product shots on a clean background, portraits with soft light, and landscapes with layers (foreground, middle, sky) all give the maker room to move the camera convincingly.
Very busy collages, heavy text overlays and tiny subjects tend to animate poorly. If you have a choice, upload the cleanest version of the image at the largest size you have.
Describing motion that looks natural
Describe one or two movements, not ten. "Slow orbit to the right, light glinting on the metal" is enough for a product. "Gentle breeze in the hair, subtle smile, camera holds still" is enough for a portrait. For scenes, name what in the environment should move: water, clouds, steam, leaves, passing traffic.
If you want the camera to stay put and only the subject to move, say so. If you want a dramatic push-in or pull-out, name the speed. The maker follows plain instructions well.
First-and-last-frame tricks
Uploading a second image as the last frame tells the maker where to end. Use it for a packaging reveal (closed box to open box), a colour-change, a room before and after staging, or a product rotating a quarter turn between two photographed angles.
Keep both images at the same aspect ratio and similar framing for the smoothest result.