dreamrec

dreamrec / UniVidX_ComfyUI

Public

UniVidX intrinsic & alpha video decomposition custom nodes for ComfyUI

houyuanchen111.github.ioUniVidX.github.io alpha-matting comfyui comfyui-custom-nodes intrinsic-decomposition siggraph-2026
46
2
100% credibility
Found May 13, 2026 at 46 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

ComfyUI custom nodes for running UniVidX to decompose video clips into intrinsic layers (RGB, albedo, irradiance, normal) or alpha components (composite RGB, matte, foreground, background) for VFX and creative workflows.

How It Works

1
🔍 Discover video magic

You stumble upon this ComfyUI add-on that breaks down everyday videos into hidden layers like pure color, lighting, shapes, or clean cutouts.

2
💻 Check your setup

See if your powerful computer with a beefy graphics card (24GB+ memory) is ready for heavy lifting.

3
📂 Slip it into ComfyUI

Drop the files into your ComfyUI folder and run a quick setup helper to get everything connected.

4
⬇️ Fetch the big brains

Download the huge thinking files from a safe online spot so the tool knows how to analyze videos.

5
✨ Run your first breakdown

Drag a sample plan onto the canvas, pick your video clip, choose a magic mode, and press play – it chugs away creating layers!

🎉 Unlock creative powers

Celebrate as you get separate video tracks for relighting scenes, pulling perfect cutouts, or feeding into 3D tools without green screens.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 46 to 46 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is UniVidX_ComfyUI?

UniVidX_ComfyUI brings UniVidX video decomposition to ComfyUI via Python custom nodes. It decomposes clips into intrinsic layers like RGB, albedo, irradiance, and normals, or alpha components including mattes, foreground, and background—30 modes total from a simple five-node graph. Users get physically grounded outputs for relighting or VFX without scripting pipelines.

Why is it gaining traction?

Unlike generic Wan2.1 wrappers, it specializes in UniVidX's decomposition tasks no one else supports, with drag-and-drop ComfyUI workflows and FP8 optimization slashing VRAM to 14GB and wall times by 13% on RTX 5090. Benchmarks, chunked samplers for long videos, and presets for production or previews hook power users processing multi-hour clips overnight.

Who should use this?

VFX pros pulling clean alpha mattes from raw footage, 3D pipeline devs feeding normals/albedo into reconstruction tools, or ComfyUI video artists building relighting/ControlNet chains. Needs 24GB+ VRAM GPUs like RTX 4090/5090; skip if you're on lighter hardware.

Verdict

Solid for high-end ComfyUI video decomposition if you match the hardware—docs shine with per-GPU tables and workflows, plus 85 unit tests. At 46 stars and 1.0% credibility, it's early but validated; test on your rig before committing 85GB disk.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.