Stable Video Diffusion

You give Stable Video Diffusion a still image and it returns a short video built from that frame. That is the whole workflow, and it is why the model suits developers who want to experiment with image-to-video generation.

Stable Video Diffusion is an open-source AI video generator from Stability AI, so you can work with it directly instead of going through a closed service.

The catch is control.

Text prompts cannot steer it, so you get what the image suggests rather than what you describe.

Some outputs have little or no motion, text often renders illegibly, and faces and people can come out inaccurate.

It is also not intended for commercial use at this point.

Use it as a research and prototyping model for animating stills, with high-quality results on the right input.