Get Lora Stack Extra

Get Lora Stack Extra reads a stored LoRA stack out of an image's image_info extras and returns it as a proper IPT-LoraStack socket. If you've ever loaded a generated image and wished it could just tell you which LoRAs and weights were used - this node is the mechanism by which that wish gets granted, inside this pack's ecosystem.

The bigger picture: LoRA reuse is the payoff

LoRA handling in ComfyUI is where workflow reuse usually dies. You've got a stack of three LoRAs, each with a weight that took you half an hour of fiddling, and the only record of it is the screenshot you took. This pack attacks that from both ends: Image Info Context extracts <lora:name:weight> tags straight out of the positive prompt, the pack's metadata round-trip keeps the stack attached to the image file, and the Set/Get extras give you an explicit place to park a stack under a named key.

Get Lora Stack Extra is the read side of that named-key storage. A Set Lora Stack Extra writes the stack into image_info extras, this node pulls it back out, and the resulting IPT-LoraStack plugs into the pack's loader stack (or anything that accepts the pack's stack type). Because extras ride along in saved metadata, the stack follows the image rather than the workflow.

Inputs and output

Same two-input shape as the rest of the family:

• extras - the IPT-ImageInfoExtras socket.
• key - the parameter-line key to look up.

Output is a single lora_stack. Missing key means None - and note the pack's convention: an explicit empty list means "no LoRA" and is preserved, whereas None means "not set." If you're composing fallbacks, that distinction matters.

The honest caveats

First, LoRAs are only reusable if the model files are actually on the machine doing the loading - metadata stores references, not weights, and a missing file fails the same way it fails everywhere else. Second, weight semantics vary by architecture: a stack tuned for a Pony or Illustrious base is not automatically sensible on SDXL or Flux, and the "LoRA not working" section of the community troubleshooting playbook is mostly this. The stack is a record of your intent, not a guarantee of compatibility.

Installing it

Ships in the ComfyUI-Info-Prompt-Toolkit pack. Via ComfyUI Manager - search "Info Prompt Toolkit" - or:

cd ComfyUI/custom_nodes
git clone https://github.com/kinorax/comfyui-info-prompt-toolkit.git
cd comfyui-info-prompt-toolkit
pip install -r requirements.txt

Restart ComfyUI. No model downloads for this node. Small single-author pack (GPL-3.0) with a modest footprint - its LoRA-in-metadata pattern is genuinely ahead of the pile of "which LoRA was this?" posts.