Object Quick Fur

Object Quick Fur is Blender's one-click fur generator, exposed as a ComfyUI node. In real Blender it's the "Add Quick Fur" button that drops a hair particle system onto your mesh in one click instead of a half-hour of panel fiddling. In Avatar Graph it does the same thing inside the headless Blender scene, which makes it the fastest route to a furry or haired avatar if you're building something beyond the standard blink-and-lipsync face.

How it works

It wraps bpy.ops.object.quick_fur, which adds a hair particle system plus a couple of modifiers to the active mesh. The settings map straight to the operator:

• BPY_OBJ - the mesh to cover in fur.
• density - LOW / MEDIUM / HIGH, how many hair strands.
• length - strand length, default 0.1.
• radius - hair radius, default 0.001. Crank this up if the fur looks invisible or threadbare.
• view_percentage - how much of the hair shows in the viewport, default 1.0. This is a performance dial, not a look dial: lower it while you iterate, keep it at 1 for final export.
• apply_hair_guides, use_noise, use_frizz - all on by default. Noise and frizz give strands organic waviness; guides help control the groom. Turn them off if you want clean, stiff hair.

The single output is BPY_OBJ, the same furry object, so it keeps flowing into Join Meshes or wherever the rest of your graph goes.

Where it actually fits

This is a "quick" tool by design - great for a shaggy dog avatar or a wild-haired character, and it lives in the demo era of the pack (the README literally warns that nodes and templates are still changing). It is not a production groom. If you need a specific hairstyle, you'll be fighting particle settings that were built for speed, not art direction. But for "make it fluffy," it's genuinely one click.

Installing it

Same pack as the rest of Avatar Graph. ComfyUI Manager → search "Avatar Graph" → install → restart, or by hand:

cd ComfyUI/custom_nodes
git clone https://github.com/avatechai/avatar-graph-comfyui
cd avatar-graph-comfyui
python -m pip install -r requirements.txt

And the pack-wide gotchas: bpy==3.6.0 pins you to Python 3.10, and the first import auto-downloads a ~2.5GB SAM checkpoint plus mediapipe landmarker files, so budget for a slow first boot. Fur also adds particle-simulation data that can slow the export; if your GLB comes out heavy, that's usually the hair, not your mesh.