petmycat

petmycat / ComfyUI-gen2

Public

Some custom nodes made for ComfyUI to accommodate things like QwenImage 2512's ControlNet Union Fun model released by AlibabaPAI.

20
1
100% credibility
Found Feb 05, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Custom nodes for ComfyUI that integrate QwenImage ControlNet for image generation with full compatibility to VideoX-Fun pipelines, plus enhanced pose detection utilities.

How It Works

1
🔍 Discover the Helper Pack

You find this handy add-on while looking for ways to make your AI art tool smarter at handling human poses and special image effects.

2
📦 Prepare Your Art Studio

You make sure your main AI image creator and a few supporting packs are all set up and ready to go.

3
⬇️ Add the Extension

You download the pack and slip it into your AI tool's extras folder so it recognizes the new features.

4
💾 Download Helper Files

You grab a folder of special language files from a trusted spot and place them where your tool expects them.

5
✨ Open Example Creation

You load a ready-made example setup, and your new pose tools and image controls come alive on screen.

6
🎨 Guide Poses and Effects

You pick images, detect body positions with adjustable confidence, and apply magical controls to shape your artwork.

🎉 See Stunning Creations

Your AI tool produces incredible images or videos that match exactly what you pictured, full of precise poses and styles.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 20 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 ComfyUI-gen2?

ComfyUI-gen2 delivers custom Python nodes for ComfyUI, letting you run AlibabaPAI's QwenImage 2512's ControlNet Union Fun model with pixel-perfect output matching VideoX-Fun's diffusers pipeline. It bridges ComfyUI's optimized model loading—handling fp8, GGUF quantized weights—for base models, VAEs, and CLIPs, while adding dedicated nodes for ControlNet loading, text encoding, LoRA merging, and sampling. Developers get drop-in workflows for video generation tasks, plus a DWpose utility with adjustable confidence thresholds for body, hand, and face keypoints.

Why is it gaining traction?

It stands out by guaranteeing identical results to VideoX-Fun without ditching ComfyUI's efficient loaders or native sampler integration, saving memory and load times on quantized models. The included example workflows and tokenizer setup make prototyping QwenImage 2512's fast, even as some GitHub repos like this one have checks that haven't completed yet or comments not visible on classic views. For ComfyUI power users eyeing some GitHub projects for video ControlNets, the exact compatibility hooks them in.

Who should use this?

ComfyUI workflow builders generating AI videos with pose control or inpainting via QwenImage 2512's models. Video AI experimenters matching VideoX-Fun outputs precisely, or pose detection users needing threshold-tuned DWpose for cleaner keypoints in custom pipelines. Skip if you're not deep into ComfyUI or VideoX some GitHub repos.

Verdict

Grab it if QwenImage 2512's ControlNet is your target—solid docs, workflows, and Apache 2.0 license make setup straightforward despite 19 stars and 1.0% credibility score signaling early maturity with no tests. Test in a fresh ComfyUI install first; it's niche but nails the promise for compatible video gen.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.