WarpFusion
Experimental alpha-masked diffusion tool for AI video effects.
About
Stable WarpFusion is an experimental AI video tool by independent developer Sxela, distributed to Patreon supporters. It applies Stable Diffusion-based effects to existing video, using alpha-masked diffusion for precise control over which regions of each frame get transformed.
It includes Face ControlNet and Temporalnet modules for facial consistency and temporal coherence, Tiled VAE, pattern replacement for prompt filtering, and settings stored in frame EXIF data. It runs in Google Colab or locally on NVIDIA GPUs with 8GB+ VRAM. It is explicitly not production-ready software and suits creators comfortable with nightly-build AI workflows.
What it can do
- Alpha-masked diffusion video effects
- Face ControlNet and Temporalnet modules
- Prompt filtering with pattern replacement
- Settings stored in frame EXIF data
- Colab and local GPU workflows
Setup
Support the developer on Patreon for notebook access, then run in Colab or locally.
Prerequisites
NVIDIA GPU with 8GB+ VRAM or Google Colab Pro, plus a Patreon subscription.
Use cases
Works with
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 120; Trending score 68/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.