AI product videos: turn product photos into short videos, frame animations and video edits
Generate ecommerce short videos from product photos or reference clips in Zalify Canvas: Reference, Frames, Edit and Extend modes, priced per second, with Seedance, Veo and Kling models.
In Zalify Canvas, wire a product photo into a Video prompt card, describe the shot and the action, and AI makes a short product video. Videos are saved to your Assets library automatically, ready to download for ads and social media.
Make a product video
Select the product photo and add a Video prompt from the + on its right edge.
Choose a mode at the top of the card (see below) and wire in what it needs.
Describe the action and the camera — for example "the camera circles the product slowly as water drops slide down the bottle, soft morning light".
Choose the model, aspect ratio, quality and length, check the estimate on the button, then choose Generate video.
A video usually takes a few minutes. You can keep working meanwhile, and closing the page doesn't stop it.
Four modes
| Mode | Use it to | What to wire in |
|---|---|---|
| Reference | Have a product or person perform what you describe, optionally following another video's camera work | Product, person or reference videos — up to 4 |
| Frames | Move from one picture to another — great for turning storyboard frames into shots | A start image; the end image is optional |
| Edit | Change something in an existing video, such as the background | The video to edit, plus reference images if you like |
| Extend | Continue an existing video forwards (or backwards) | The video to extend, plus reference images if you like |
In modes like Reference, each reference gets its own row on the card, numbered "Image 1", "Video 1". Choose Insert to cite it in your prompt — "the model from Image 1 dances like Video 1".
Choose a model
| Model | Modes | Length |
|---|---|---|
| Seedance 2.5 | All four | 4–30 s |
| Seedance 2.0 | All four | 4–15 s |
| Veo 3.1 / Veo 3.1 Fast | Frames | 4, 6 or 8 s |
| Kling 3.0 | Frames | 3–15 s |
For Reference, Edit or Extend, choose a Seedance model.
When there's a person in the video
Video models can mistake an AI-generated person for a real one and refuse to render. Register the person as an avatar and they can be used in video — with the same face in every shot. See Avatars, characters and scenes.
Pricing
Video is charged by length: the per-second price of the quality you pick, times the number of seconds.
| Model | Price per second |
|---|---|
| Seedance 2.5 | from US$0.143 |
| Seedance 2.0 | from US$0.091 |
| Veo 3.1 | US$0.52 |
| Veo 3.1 Fast | US$0.195 |
| Kling 3.0 | US$0.3276 |
More tiers and example costs are on AI model pricing. Only successful videos are charged.
Save money: settle the look with images first, try the motion at a lower quality or a shorter length, then make the final version.
Frequently asked questions
Can I make a video from a single product photo?
Yes. Choose Reference or Frames, wire in the photo, and describe the action and camera.
How much does a product video cost?
It's charged per second — a 5-second video on Seedance 2.0 starts at about US$0.46. The actual cost depends on the model, quality and length, and the button shows the estimate first.
How long can an AI video be?
Up to 30 seconds per video, with Seedance 2.5.
Why can't I choose some modes?
Veo and Kling support Frames only. For Reference, Edit or Extend, switch to a Seedance model.
A person in my video was blocked. What now?
Register the person as an avatar; once approved, pick them under My avatars on the video card.