TemporalKit is an open-source extension for AUTOMATIC1111’s Stable Diffusion WebUI that adds temporal stability to AI-generated video renders. With around 2,000 stars on GitHub, it became a popular tool in the early AI video community for turning flickery frame-by-frame generations into smooth, coherent clips.
Generating video by running Stable Diffusion on each frame independently produces distracting flicker, because every frame is drawn slightly differently. TemporalKit solves this with an EbSynth-based workflow: it extracts keyframes at a chosen interval, processes them with Stable Diffusion, and then propagates those styled keyframes across the in-between frames so motion stays consistent. The extension handles the whole pipeline inside the WebUI, including video splitting, batch keyframe processing and final reassembly, and only requires FFmpeg installed on the system. Settings such as FPS, batch size and frames per side let users trade quality against processing time. The repository was last updated in 2024.
Because the whole workflow lives inside the familiar Stable Diffusion WebUI, creators do not need to learn a separate video application: footage is imported, keyframes are generated with existing checkpoints and settings, and the stabilized result is rendered back out as a finished clip.
TemporalKit is completely free and open source under the GPL-3.0 license. It can be installed from GitHub into any local Stable Diffusion WebUI setup at no cost, with no accounts, subscriptions or hosted fees involved.
TemporalKit is for AI video creators, animators and Stable Diffusion enthusiasts who generate video locally and need watchable, flicker-free results. It fits hobbyist film experiments, music videos, and stylized animation tests built on the AUTOMATIC1111 ecosystem.