Mesh Customdata Bevel Weight Vertex Clear

Like most of the Mesh_Customdata* nodes in Avatar Graph, this one has exactly one input and one output - no dropdowns, no numbers to tune. It's a direct wrap of Blender's bpy.ops.mesh.customdata_bevel_weight_vertex_clear operator, and it does exactly one thing: it removes the per-vertex bevel weight data layer from a mesh, if one exists.

What it actually does

Blender doesn't store a "vertex bevel weight" on every mesh by default - it's optional custom data that only gets allocated when you (or a node like Mesh_CustomdataBevelWeightVertexAdd) explicitly ask for it. That layer lets a Bevel modifier or bevel operation treat some vertices as "beveled more" and others "beveled less" or not at all, instead of applying one flat amount everywhere. This node is the teardown step: it deletes that layer entirely. Any per-vertex weight values you'd set are gone, and downstream operators that check for vertex-weighted beveling fall back to treating every vertex the same.

Practically, you'll reach for this far less often than the Add version. Removing custom data layers you don't need is mostly a memory/cleanliness thing - useful if you're exporting a mesh and want to strip attributes that won't be read anywhere downstream, or if an earlier step in a generated pipeline added the layer speculatively and it turned out unused.

The input and output that matter

• BPY_OBJ (optional) - the mesh object to strip the vertex bevel-weight layer from.

Output is the same BPY_OBJ, with that one data layer gone if it was present. If the mesh never had a vertex bevel-weight layer in the first place, this is a safe no-op - you won't get an error for clearing something that isn't there.

How to install it

Search Avatar Graph (avatar-graph-comfyui) in ComfyUI Manager and install from there, or do it 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

The whole pack hinges on one pinned dependency: bpy==3.6.0, Blender's Python API, which only builds against Python 3.10.x. On Windows, grab the README's prebuilt Python-3.10 ComfyUI zip rather than fighting your existing install. On Mac/Linux, a dedicated conda environment (conda create --name comfyui python=3.10) is the path of least resistance. requirements.txt also pulls in opencv-python, mediapipe, and segment-anything for the pack's segmentation side - expect a heavier install than a typical node pack. Restart with python main.py --enable-cors-header afterward if you plan to use the live avatar preview.

Common issues & troubleshooting

"Running it did nothing I can see." Correct, and expected - clearing a data layer that has no effect on shape or shading unless something downstream was actually reading vertex bevel weights. If you're trying to visually undo a beveled look, this isn't the node for that; the bevel geometry itself is created elsewhere (a Bevel modifier or a bevel operator), not by this data layer alone.

Errors on import, not on running this specific node. That almost always traces back to the Python version pin. If ComfyUI's console shows an import failure for bpy at startup, confirm you're actually running the 3.10.x environment the pack expects - a system Python 3.11 or 3.12 install will not load bpy==3.6.0, full stop.

Wondering if you need this at all. Honestly, most Avatar Graph workflows built around the character-rigging templates in the README never touch vertex bevel weights, added or cleared - this node exists because the pack auto-generates a ComfyUI wrapper for a large chunk of Blender's mesh operator surface (see generate_blender_types.py in the repo), not because it's a load-bearing part of the avatar pipeline. If you didn't add the layer with its counterpart node, you don't need to clear it either.