Photoshop Add Layer

The endgame of the live loop: instead of eyeballing renders in a preview panel and then manually recreating them in your document, Photoshop Add Layer takes an image and pushes it straight into the Photoshop document you're working on as a new layer, scaled to the document bounds (pixels capped at 4096px on the long side). It's the "commit the good one" step - the difference between "AI made a nice face, now I hand-copy it" and "AI made a nice face, it's a layer in my document."

It's a pure endpoint: one required image input, no outputs. The only widget is layer_name (default "ComfyUI refined") - the name the new layer gets in Photoshop, worth setting if you're stacking several.

What it's for, and what it isn't

This node exists to end a refine chain: render something good → run the high-quality refine pass → Photoshop Add Layer puts the result into your document. The panel's REFINED LAYER control decides whether repeated refines Stack new layers or Replace the previous one - Stack is handy for comparing candidates side by side, Replace keeps the document tidy when you're dialing in one image.

It is not the same thing as ComfyUI 0.31+'s own core Add Layer node, and the pack is careful to say so: core's Add Layer appends a layer to a LAYERS stack inside the graph. This one delivers pixels to the live Photoshop document. Different jobs, same name, easy to trip over when you're googling.

One more honest note: without the plugin connected, this node runs and logs a no-op rather than failing your render - the image's already visible in ComfyUI either way, and the pack never wants a dead-end node to kill a workflow. But "delivers into Photoshop" is the whole point, so install the Tier 2 plugin if you actually want layers to land.

Install

Same as every node in this pack - Manager → Install via Git URL, or:

cd ComfyUI/custom_nodes
git clone https://github.com/ericpaulsnowden/comfyui-photoshop-bridge.git comfyui-photoshop-bridge
pip install -r comfyui-photoshop-bridge/requirements.txt

…with ComfyUI's own Python, then restart. Pre-1.0, actively developed. Practical tip: if your refined layer lands slightly off or soft, it's usually the capture-to-document scale math - verify your canvas and capture sizes line up, and keep the document's long side under 4096 to avoid the cap kicking in.