PNG Info — Быстрое применение

If PNG Info for ComfyUI is the full kitchen, this is the drive-through window. PNG Info - Быстрое применение ("Quick Apply") is a tiny companion node from the same pack that does one thing: upload a PNG, and re-apply its metadata through a binding you already saved in the main node.

The main PNGInfoForComfyUI node has you bind a five-node target set - checkpoint loader, positive/negative CLIP text encodes, KSampler, LoRA loader - then detach all the temporary wires. That setup is the powerful part, but you don't want to redo it every time you drag in a fresh PNG. This node exists so you don't have to: keep one fully-configured main node in the workflow, park this button-only node beside it, and everyday reuse becomes load-an-image, click-apply.

How it works

Look at its inputs and outputs and there's nothing there - zero sockets on both sides. That's not a bug. The backend class is a no-op output node whose function just returns (), and all the real work happens in the pack's frontend extension (web/png_info.js). The node stores a controller ID in the workflow, which the extension uses to find the main PNG Info node and its saved binding. When you hit apply it reads the PNG, parses it with the same backend parser the main node uses, and pushes the values into the bound KSampler branch.

The interface is deliberately bare: a Загрузить картинку (load image) button and a Применить PNG (apply PNG) button. That's the whole node. It doesn't expose prompts, seeds, or settings to edit - if you want to fiddle, you edit the widgets on the bound main nodes afterward, or use the full node.

Setting it up

You don't install anything extra - the node ships in the same pack, so comfy node install png-info, a ComfyUI Manager search for "PNG Info for ComfyUI", or a manual clone of https://github.com/Gerbesh/png-info-for-comfyui into custom_nodes brings it along, then restart ComfyUI. It shows up as "PNG Info - Быстрое применение" under the PNG Info category.

Before the quick node will do anything, a configured main node has to exist in the graph with a saved binding. No binding, no apply. And if you have several configured main nodes in the workflow, the compact node refuses to guess which one you meant - the README is explicit about it refusing. That's a feature, honestly: silently applying an old PNG to the wrong branch would be worse than a refusal you can fix in ten seconds by removing or unbinding the extras.

When to reach for it

This is the node for the "I have a folder of 500 A1111 PNGs and I'm checking how each one looks in this new model" workflow, where binding once and clicking through the stack beats reconnecting wires per image. For one-off work you'll probably just use the main node directly - but once you've bound a branch you actually like, the quick node is the part you end up clicking every day. There are only two buttons on it. That's the point.