Pre-baked ComfyUI pipelines for QR spirals and face posing

What it does This repository is essentially a curated recipe box for ComfyUI. It hosts dozens of workflow images—covering SD 1.5, SDXL, SSD-1B, Stable Video Diffusion, and WAN—that encode complete node graphs. The author also maintains a resource table linking to the required custom nodes, ControlNet models, and IP Adapter checkpoints needed to run them.

The interesting bit The workflows target narrow, specific effects: QR-code spirals, “Instant LoRA” face mimicry without training, AnimateDiff video morphing, and a face reposer that deliberately avoids roop-style face-swap nodes. The README is unusually candid, flagging deprecated nodes and admitting some experiments “need more work.”

Key highlights

• ~30+ workflow recipes ranging from basic SDXL ControlNet to WAN Fun Control
• Workflows are embedded in PNG metadata that ComfyUI reads natively
• Covers video generation (AnimateDiff, SVD), image restyling, and upscaling pipelines
• Includes a resource index for models, ControlNets, and IP Adapter checkpoints
• Explicit deprecation notes for outdated custom nodes (e.g., 2023 IP Adapter implementations)

Caveats

• This is documentation and curation, not a standalone application or library
• Several workflows rely on deprecated custom nodes that require manual replacement
• The author notes that some experiments, like the MagicAnimate workflow, are incomplete

Verdict Worth bookmarking if you already run ComfyUI locally and want to bypass node-wrangling for specific visual tricks. Not for you if you need a turnkey installer or a maintained software package; this is a reference shelf that expects you to hunt down your own models.