Lora List Stacker

Instead of wiring up one LoRA-loader node per LoRA - the usual daisy-chain that turns into spaghetti fast - this builds a whole LORA_STACK from a single text field. You list your LoRAs and their strengths as text, and the node emits a stack that a stack-applier node then applies to your model and CLIP in one shot. If you've used the efficiency-nodes or rgthree "Lora Stacker" pattern, this is Art Venture's take on it, with the list living in text so it's easy to template, paste, or drive from an API.

The LORA_STACK type is the key. It's a compact list of (lora, strengths) entries that travels down one wire, and because it's data, you can assemble it programmatically, swap the whole set by editing one text box, and - via this node's chaining input - stitch several stackers together. That's the appeal over hand-wiring loaders: your LoRA config becomes one editable thing.

How it works

It reads the data text, parses each entry into a LoRA name plus its strength(s), and outputs a LORA_STACK. If you pass an existing stack into the optional input, it appends to it rather than replacing - that's how you compose stacks across multiple nodes. Nothing is applied here; the stack is a spec that a downstream "apply LoRA stack" node turns into actual model/CLIP patches.

The inputs and outputs that matter

• data - a multiline text field listing the LoRAs and their strengths. Heads-up: the exact line syntax isn't spelled out in the pack README, so if a LoRA isn't taking effect, check the node's own tooltip/placeholder in the UI and match its format precisely (LoRA name as it appears in your models/loras folder, plus the strength). Get the name wrong and that line is silently skipped.
• lora_stack (optional) - an incoming LORA_STACK to build on top of. Wire another stacker (or this node chained to itself) in here to combine multiple lists into one.

Output is a single LORA_STACK. Feed it into a stack-applier node (the one that takes a LORA_STACK and your MODEL/CLIP and returns patched versions).

How to install it

Part of the Art Venture pack. ComfyUI Manager → search comfyui-art-venture → Install → restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/sipherxyz/comfyui-art-venture

then pip install -r comfyui-art-venture/requirements.txt and restart. No model download for the node itself - you supply the LoRA files in ComfyUI/models/loras.

Common issues & troubleshooting

A LoRA in my list does nothing. Two prime suspects: the name in data doesn't exactly match the filename in models/loras (extensions, subfolders, and casing matter), or the line syntax is off. A malformed or unmatched entry gets skipped without a loud error, so it looks like the LoRA "didn't apply" when really the parser never found it. Fix the name/format and it'll take.

The stack output isn't applying to my image. Remember this node only builds the stack - it doesn't patch your model. You need a stack-applier node after it, with your MODEL and CLIP running through that applier. If you wired the LORA_STACK straight past the applier, nothing happens.

Chained stacks aren't combining. Make sure you're feeding the upstream LORA_STACK into this node's optional lora_stack input (which appends), not into a second independent applier. The chaining is what merges lists.

Node missing after a workflow import. The pack isn't installed - Manager → Install Missing Custom Nodes.