Turn a photo into a short video with Wan 2.2, Alibaba's open-weights image-to-video model: upload an image, describe the motion you want, and get a clip of about five seconds that keeps the subject and style of the picture. Free, no account, running in the embedded demo below.
Wan 2.2 is the successor to Wan 2.1 in Alibaba's line of open video models, released in July 2025 under the Apache 2.0 licence, with text-to-video and image-to-video variants that use a mixture-of-experts architecture for better motion and detail than Wan 2.1.
The model itself produces 480p or 720p clips of about five seconds at 16 frames per second. The embedded demo is hosted by a third party and sets its own limits, usually 480p; they are shown in its interface.
Video generation is heavy: a single clip needs a large GPU for a minute or more, and the queue is shared. This demo is hosted by a third party (Hugging Face Space) and embedded here; it can be slow, queued or temporarily unavailable when the host is busy. Nothing to install, no account needed.
Read the guide: How to write a prompt: the same rules apply to motion prompts