AnimateDiff
Open-source AI tool that adds motion to Stable Diffusion images to produce short animated video clips from text or image inputs.
About
AnimateDiff is an open-source research project that extends Stable Diffusion with learned motion modules, enabling it to generate short animated video clips rather than static images. You provide a text prompt or a starting image and the model produces a sequence of frames with smooth, coherent motion. It supports looping animations, image-to-video conversion, and video-to-video editing when combined with ControlNet. AnimateDiff runs locally and requires an Nvidia GPU with at least 8 GB of VRAM, and it integrates with popular interfaces such as Stable Diffusion WebUI and ComfyUI. Third-party sites also offer free online access without local hardware. The project is entirely free and open-source.