AI Image to Video
Upload a photo, describe the motion, get a clip — two Wan image-to-video models, priced per 5 seconds.
Image to video keeps the composition you already have and adds motion to it. You upload a still — a product shot, a portrait, an illustration — and the model animates the subject, the camera, or both. It is the fastest route to consistent output, because the first frame is set by you rather than guessed by the model.
Your frame stays your frame
The uploaded image becomes the first frame, so composition, colour and product geometry are locked before generation starts.
Cheapest route to motion
From 30 credits for a 5-second 480p clip — well under the price of a comparable text-to-video render.
Silent by design, priced accordingly
Both image-to-video models render without audio, which is why they cost less than the audio-capable Wan 2.7.
Compare video models
| Model | Resolution | Video length | Price |
|---|---|---|---|
| Wan 2.6 I2V · Silent | 720p / 1080p | 2s – 15s (any whole number of seconds) | 40 credit / 5s |
| Wan 2.2 I2V · Silent | 480p | 5s (fixed duration) | 30 credit / 5s |
How to make one, start to finish
- Pick a source image with a clear subject — One subject, uncluttered background, decent light. The model animates what it can see — a blurry or crowded frame gives it nothing to work with.
- Describe the movement only — The image already decided what is in the frame. Your prompt should say what moves: a slow push-in, hair drifting left, steam rising from the cup.
- Choose resolution and length — Wan 2.2 I2V is fixed at 5 seconds and 480p. Wan 2.6 I2V goes to 1080p with any whole-second length between 2 and 15.
What creators make with it
- Product shots turned into loopable 5-second clips for listings and ads.
- Portraits and character art brought to life with subtle head or camera movement.
- Illustrations animated for explainer videos and social posts.
- Architectural and interior stills with a controlled camera push.
FAQ
Which models support image to video?
Two: Wan 2.6 I2V (480p to 1080p, 2 to 15 seconds) and Wan 2.2 I2V (480p, fixed 5 seconds). Both are silent — they render video only, without a generated soundtrack.
Why are the image-to-video models silent?
Both upstream models are the no-audio variants on Alibaba's DashScope platform. Rendering without audio keeps the per-second cost at roughly half of the audio-capable model, which is why they are the cheapest video options on Slloy. Add music in your editor afterwards.
What image should I upload?
A single clear subject, preferably 720p or above, separated from the background. Avoid heavy compression artefacts and frames with many competing elements — the model animates whatever is in the frame.
Can I control how much things move?
Indirectly, through the prompt. Words like subtle, slow or slight breeze reduce motion; fast pan or explosive increase it. Very large motion from a single still tends to warp the subject, so keep moves modest.
Do the videos have a watermark?
No. Output from Slloy is delivered without a watermark on every plan, including the free tier.