Blender Render Image

Everything in Avatar Graph's Blender side lives as BPY_OBJ - live objects inside a headless Blender scene you can't actually see unless you're running with --enable-cors-header and watching the live preview. Blender Render Image is the bridge back to something ComfyUI (and you) can look at directly: it renders whatever's in the scene and hands you back a normal IMAGE tensor, so the rest of your graph - a save node, a mask, an upscaler, anything - can treat it like any other image.

How it works

Under the hood this is Blender doing what Blender always does - rendering the current scene from its camera through whatever render engine and settings are configured - but triggered from a ComfyUI node instead of Blender's own UI. That makes it a useful checkpoint node in the middle of a longer graph: after you've built and rigged some mesh geometry with the pack's other Blender nodes, you can drop this in to actually see what you've built, without leaving ComfyUI or exporting a file. It's also the node you'd reach for if you want to feed a rendered pass - say, a normal map or an AO bake from Object Bake applied to the mesh - back into an image-side pipeline for further processing.

The inputs and outputs that matter

There's exactly one input, and it's optional at the schema level but functionally required for a useful render:

• BPY_OBJ - the Blender object (or scene context) to render. Chain this from whatever mesh-building node produced the thing you want to look at.

Two outputs come back:

• BPY_OBJ - passes the object through, so you can keep chaining Blender operators after rendering without breaking the graph.
• IMAGE - the actual rendered picture, ready to wire into anything that expects a standard ComfyUI image.

How to install it

ComfyUI Manager, search "Avatar Graph", install, restart - or manually:

cd ComfyUI/custom_nodes
git clone https://github.com/avatechai/avatar-graph-comfyui

then pip install -r requirements.txt and restart. This node needs Blender's bpy module, which is pinned to Python 3.10.x specifically - on macOS/Linux, run the whole pack inside its own conda environment built for that version rather than an existing newer ComfyUI setup. Windows users have it easier: Avatech provides a prebuilt ComfyUI + Python 3.10 zip, then you install this pack through Manager on top of it.

Common issues & troubleshooting

If the node loads but the rendered image comes back black, empty, or clearly wrong, the most likely cause isn't this node - it's that nothing meaningful was actually in the scene when it fired, or the camera/lighting the render depends on wasn't set up by whatever's upstream in your graph. Since this pack chains Blender objects node-to-node rather than giving you Blender's normal viewport to check your work as you go, it's worth inserting a BlenderRenderImage checkpoint earlier in a complex graph rather than only at the end, so you can catch a broken mesh-building step before it propagates.

If the node fails to load entirely, that's the pack's bpy dependency, not a render-specific bug - check ComfyUI's startup console for an import error and confirm you're running Python 3.10.x with pip install -r requirements.txt completed cleanly. This class of "wrong Python version breaks the whole pack" issue is common to any ComfyUI extension that embeds Blender's Python module, and it's the first thing to rule out before assuming a specific node is broken.