VRGDG_ShowImage

VRGDG_ShowImage is as close to nothing as a node can be: one optional image input, no outputs, and it saves whatever image reaches it to your ComfyUI output directory. The official description says it exactly - "Saves the input images to your ComfyUI output directory." It's a terminal node, meaning it exists to be the end of a wire, and its whole purpose is making sure an intermediate image actually gets written to disk where you can find it.

Why you'd use a node that does so little

In ComfyUI, saving images isn't automatic. Outputs only hit disk if a save node is on the wire, and most workflows only save the final result. When you're debugging or staging a long pipeline - like the multi-scene video renders this pack is built around - you often want to keep intermediate frames: a storyboard frame, a reference frame after preprocessing, a single extracted frame from a video. That's this node. Wire an image into it and it lands in your output folder, no filename ceremony, no format questions.

The "Show" in the name is doing work: as an output node, it also renders the image in the ComfyUI UI when you run, which is the fastest way to see a frame mid-pipeline. So it's both a save-to-disk and a look-at-this node in one.

Where it fits

The music-video workflow runs long and mostly unattended. A single ShowImage parked after your storyboard image step means when you come back after an hour, there's a frame on disk showing you exactly what the storyboard stage produced - before video generation touched it. Same trick works after a VAE decode, after a frame extraction, anywhere you want a checkpoint you can open in an image viewer without hunting through ComfyUI's temp files.

It's also the polite way to end a debug wire. ShowAny gives you text; ShowImage gives you pixels. Put them together and there's very little in a graph you can't inspect.

Install

It rides along with the pack, so:

cd ComfyUI/custom_nodes
git clone https://github.com/vrgamegirl19/comfyui-vrgamedevgirl

or ComfyUI Manager → "vrgamedev", then restart. Requirements: kornia, librosa, imageio (on Windows portable, install from python_embeded).

Honest sizing: this is a utility node and proud of it. It won't wow you, and it's not meant to. But when you're deep in a workflow you don't fully understand and you need eyes on an intermediate image, "wire it into VRGDG_ShowImage" beats "figure out where this stage saves" every time.