Photo to Video
Turn a still photo into a short clip — a slow push in, a drift across the frame, a gentle parallax.
This one takes a while. Video generation runs for considerably longer than an image — expect to wait, and keep the tab open. The page will tell you when it is done.
Your photo
It becomes the first frame. Resized in your browser before anything is sent.
What should it look like?
These choices decide the result. You never have to write a prompt.
A short note about the motion or mood — “with drifting mist”, “leaves moving in the wind”.
0/200Your image
Generation runs on our server, because it needs a model your browser cannot. Your photo is sent for this step only and is not kept afterwards.
Add an image above to enable this.
This is the workflow where your photo actually survives
Worth saying plainly, because the image workflows on this site do not work that way.
The model animates from your actual photograph rather than from a description of it. What you uploaded is what moves — the subject, the colours and the composition are yours.
Every frame after the first is generated. Faces, hands and text drift as the clip runs, and the further it goes the more it departs. Shorter clips hold together better than long ones.
What makes a photo work
- Works well
Landscapes, interiors, still life, architecture — anything with depth and no faces. Parallax has something to work with and nothing to get wrong.
- Works poorly
Close-up portraits. Faces are where generated motion is least forgiving, and small distortions read as uncanny immediately.
- Will not survive
Signage, logos and printed text anywhere in frame. They are stable in the first frame and degrade from there, which is more distracting than no motion at all.
What next
- Trim the start or endTrim Video
- It is too large to uploadCompress Video
- Add music under itAdd Music to Video
- Add a captionAdd Text to Video
Frequently asked questions
Considerably longer than an image — generation runs on the provider's side and a few seconds of video is a great deal more work than a single frame. The page polls in the background and updates when it is finished, so leave the tab open.
Yes, for the generation step. It is resized in your browser first, then sent so it can be used as the opening frame. It is not stored beyond producing your clip, and the result expires from our storage rather than being kept.
Because every frame after the first is generated, and faces are where small errors are most visible. Use it on scenes rather than portraits, and prefer shorter clips — the longer it runs the further it drifts from what you gave it.
Not from this workflow. The clip comes back silent, which is usually what you want for something going under a caption or a voiceover. Add a soundtrack afterwards with Add Music to Video.
No. Generation is not deterministic, so each run differs even from the same photo and settings. Download anything you want to keep.