💊 Easy LoRA Clothing Loader

Clothing LoRAs are the narrowest kind of adapter in the ecosystem - the entire point of the file is usually one outfit, sometimes plus its "removed" state. A full-featured LoRA loader is overkill for that, and remembering the trigger word and the right strength every time is exactly the busywork you don't need. Easy LoRA Clothing Loader, from Comfy Cabinet, is the pack's leanest specialized loader: two prompt blocks, nothing more.

It exists because the pack's designers organized prompts into eight fixed categories (character, clothing, expression, situation, location, lighting, and so on), and the clothing category maps to exactly two slots: clothing and no_clothing. If a clothing LoRA trains a "without the outfit" variant tag - many do - the second slot is where it lives.

How it works

Same skeleton as every other Easy LoRA node. You feed it model and clip, it applies the LoRA through ComfyUI's own LoraLoader, and it hands back the patched MODEL and CLIP plus the two prompt strings. The strength_model and strength_clip floats behave exactly like the core loader's - default 1.0, range -20 to 20, and 0/0 means pass the model through untouched, which is a handy way to disable an outfit without deleting the node.

The lora_name dropdown is filtered to show clothing-type LoRAs (it falls back to the full list if your filter would leave it empty). Select one and the node auto-populates the strength widgets and prompt blocks from the config you saved in the dashboard. For the prompt outputs themselves, a value you typed on the canvas wins; otherwise the saved config fills the slot.

What you set

Required inputs: model, clip, lora_name, strength_model, strength_clip - same as any LoRA loader. Optional: the two multiline fields clothing and no_clothing. Outputs: MODEL, CLIP, clothing, no_clothing. Wire model and clip down to your sampler, and feed the clothing strings into your text encoder - either directly or, if you're combining a character LoRA's output with this one, through a prompt stitching node.

Install and config

Standard for the pack:

cd ComfyUI/custom_nodes
git clone https://github.com/Zed93/Comfy-Cabinet.git

Restart ComfyUI (or search "Comfy Cabinet" in ComfyUI Manager). Only dependency is aiohttp, already present. Configure via the 🗄️ Easy Config button in the top menu → LoRA config tool. It syncs previews, tags and trigger words from Civitai if you add your own API key (both civitai.com and civitai.red are supported), caches them in civitai_cache/, and pushes saved configs straight to the matching node on your canvas.

Where people get burned

The usual trio applies: configs are keyed by exact LoRA filename, so renaming a file orphans your saved trigger and strength; the Civitai sync needs a key and generation works fine without it; and it's release-candidate software, so restart before you debug if a button vanishes. Also worth knowing which node to reach for: this one is strictly for clothing-type LoRAs. If the LoRA is really a character model whose outfit is just one aspect, the Character loader's four blocks fit better - this narrow node is the right tool when the wardrobe is the whole job.