Object Set Display Type

Every 3D package has the same four viewport display modes - solid, wireframe, textured, bounds - and Blender's object type stores which one an object is in as a plain property. ObjectSet Display Type is the Avatar Graph node that sets it. In a pack whose whole point is turning a character image into an animated mesh, this is one of the more genuinely useful setters, mostly because of what it's not doing: it changes how things look while you work, not what renders.

The default is TEXTURED, which is exactly what you want for the avatar mesh itself - you see the actual texture on the mesh as it'll appear in the final render. The value you'll actually switch to is WIRE, which turns the mesh into its wireframe so you can see the topology and, more importantly, see through it to the armature, empties, and shape-key rigging underneath. SOLID shows shaded geometry without textures (handy when you're checking a mesh's form before it's textured), and BOUNDS collapses everything to its bounding box - great for tidying helper objects that keep cluttering the preview.

How it works

Like every ObjectSet_* node, this one is auto-generated from a dump of Blender's bpy.types.Object properties (blender/input_types.txt) and, on execution, just sets that property on the object passing through - setattr, nothing clever. It's a single mutation in a chain, and the object comes out the other side unchanged so the next node can keep decorating it.

Two inputs, one output:

• BPY_OBJ - the Blender object from your mesh pipeline
• value - a string, default TEXTURED; valid values are TEXTURED, SOLID, WIRE, BOUNDS

The output is the same BPY_OBJ, ready to feed into Avatar Main Output or another setter.

Installing Avatar Graph

The node ships with the whole pack, so one install covers it:

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

Or hit ComfyUI Manager → Install Custom Nodes → search Avatar Graph. Heads-up on the heavy stuff: the pack pins bpy==3.6.0, which demands Python 3.10 (Windows users get a bundled 3.10 environment from the README; macOS/Linux should use a conda env), and the first launch downloads the SAM ViT-H segmentation model plus mediapipe face/pose landmarkers before anything shows up.

Common issues

Don't expect a wireframe object to export as a wireframe - display_type is strictly a viewport display setting, so it has zero effect on the .GLB/.GLTF you export from Avatar Main Output. People hit this all the time with Blender imports: they set WIRE to declutter, then wonder why the exported file looks different. If your goal is "don't render this helper," that's ObjectSet Hide Render's job, not this node's.