Object Set Is Embedded Data

Some Blender object flags exist to describe reality, not to change it. ObjectSet Is Embedded Data is one of those. The property it sets - is_embedded_data - records whether an object's data lives inside the current .blend file or is pulled in from somewhere else (linked libraries, appended files). It's bookkeeping, the kind of thing that answers "where did this mesh actually come from?"

Here's the honest version: for what Avatar Graph does, this node is a formality. The pack drives a headless Blender that resets to factory settings at the start of every graph run - a blank, self-contained scene with no linked libraries. Everything in it is, by construction, embedded data. So the flag is true for everything you touch, and there's nothing meaningful to change it to.

Why does the node exist at all, then? Because the whole ObjectSet_* family is auto-generated from a dump of Blender's bpy.types.Object properties, and the generator includes everything, including read-mostly flags like this one. Blender's Python API exposes is_embedded_data as read-only, which is a polite way of saying there's not much to set. Feed it a value and you're at best making the graph self-documenting.

How it works

Mechanically it's identical to its siblings: input BPY_OBJ plus a boolean value (default false), and on execution the node does a setattr on the Blender object, then returns the same BPY_OBJ for the next node in the chain.

The inputs that matter:

• BPY_OBJ - the object whose flag you're asserting
• value - boolean, default false

Output: the same BPY_OBJ. In a graph that doesn't otherwise use this node, you can leave it out entirely.

Installing Avatar Graph

This node ships with the whole pack:

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

Or ComfyUI Manager → Install Custom Nodes → search Avatar Graph → restart. The pack pins bpy==3.6.0, which requires Python 3.10 (bundled 3.10 environment on Windows via the README; conda env on macOS/Linux). First launch auto-downloads the SAM ViT-H segmentation model and the mediapipe face/pose landmarkers, so the first run is slow.

Common issues

Don't build a workflow around this flag - in the pack's fresh-scene model it's a no-op. If you ever see is_embedded_data being meaningful, it's because you're loading a pre-made .blend with linked assets into a workflow, and the flag reflects that linkage. That's not a setup Avatar Graph ships; the pack builds its Blender scene from scratch every run.