Turn a photo into AI video: upload a still, set motion strength and camera direction, and add a last frame if you want the model to fill the passage between them.
No redrawing needed.Let the still you already like start moving.
Two stills of the same scene, one medium and one wide. The camera move between them is the model to fill.


People for action, scenery for camera moves, products for texture.




Photos, illustrations and product shots all work. Clean up noise and blur first.
Strength sets how far it travels, direction sets where the frame goes.
Download if it works. If not, reuse the result as the first frame.
From almost still to big action. Talking heads stay low, action scenes go high.
Push, pull, pan, orbit. One axis at a time keeps the frame steady.
Give a start and an end still and the model fills what happens between.
9:16 without cropping, so the whole figure stays in frame.
Clear, well-lit images with a defined subject. Product shots, portraits and illustrations all work. The first frame keeps its composition, so pick an image you are happy to see in motion.
The options panel offers an end frame when the selected setup supports it. Locking both frames gives you a predictable start and finish for loops, product reveals and transitions.
If the provider rejects the job before accepting it, the credits are refunded automatically. Once the provider has accepted the job and started rendering, the run is charged.
Square, vertical and widescreen ratios are offered when the current configuration supports them. Choose vertical for short-form social and widescreen for YouTube or landing pages, and crop the source image to match before uploading.
Yes. Illustrations, renders and screenshots all work as a first frame. Keep the subject clear and avoid heavy compression artifacts, which tend to show up once the image starts moving.
Usually because the source frame is cropped badly or the movement is too strong for a single photo. Keep the subject inside the frame, reduce the motion in the prompt and regenerate.
Yes. The prompt and settings stay in place - replace the image and generate again. Reusing the same motion description across a set of product photos is what keeps a catalogue consistent rather than a collection of clips that each move differently.
Describe movement around the subject rather than of it. Camera pushes, changing light and small gestures keep a face or product in place; asking the subject to walk, turn fully around or change pose invites distortion, especially on faces and hands.
Not for judging motion. Low resolution renders faster and costs less, and you are looking at how the frame moves, not at fine texture. Move to the highest setting only for the take you plan to publish.
It depends on duration, resolution and how busy the provider queue is; the composer shows an estimate before you submit. Drafting short and at a low resolution keeps the wait down while you are still deciding on the movement.