Power Lora Loader to Prompt (Image Saver)

Two very popular nodes don't talk to each other. rgthree's Power LoRA Loader is where you stack and toggle your LoRAs - it's the single most-shared node in rgthree's pack, and for good reason. alexopus's Image Saver is what stamps metadata onto your PNGs so Civitai and other sites can auto-detect your resources. The gap: Image Saver saves LoRA info from the prompt text, and the Power LoRA Loader never puts its LoRAs in your prompt string. This vsLinx node is the bridge - it reads the LoRAs from your Power LoRA Loader and appends them to your prompt as <lora:path:strength> tokens, in exactly the format Image Saver expects.

How it works

Give it your prompt text and tell it where the Power LoRA Loader is - three ways, tried in order:

1. By link (the cleanest): wire the Power LoRA Loader's model output into the optional powerloraloader_model input.
2. By id: the node's numeric ID in the graph.
3. By node_title: the title you gave it.

It then reads which LoRAs are enabled and their strengths, and appends <lora:filepath:strength> tokens to your text. Pass that combined string to Image Saver as the positive prompt and it saves the LoRA hashes for Civitai while stripping the <lora:...> tokens from the final prompt - so your prompt stays clean and your metadata carries the loaders. There's an only_enabled toggle to skip disabled LoRAs and a debug toggle that logs resolution steps to the console when the lookup fails.

Inputs and outputs

• text (STRING, multiline) - your prompt.
• powerloraloader_model (MODEL, optional) - the link-based way to point at the loader.
• id (INT) and node_title (STRING) - the fallback lookup methods.
• only_enabled and debug - the toggles above.
• Output: a single STRING with the LoRAs appended.

Install

Part of the vsLinx pack, but it's only useful with rgthree-comfy and ComfyUI-Image-Saver installed too. All three via ComfyUI Manager, or for this pack:

cd ComfyUI/custom_nodes
git clone https://github.com/vslinx/ComfyUI-vslinx-nodes.git comfyui-vslinx-nodes

Restart. No extra dependencies beyond those two packs.

The honest take, and what actually happens in the wild

The intent behind this node is real and shared by a lot of people: get Civitai to auto-detect the LoRAs in your images. But it's worth knowing that "Civitai isn't detecting my LoRAs" is one of the most common metadata complaints in the ComfyUI community, and the advice there often ends up being "install LoRA Manager and let it handle metadata" rather than building a bridge like this. So this node is the right tool when you're already committed to the rgthree + Image Saver stack and want them to cooperate - it's the glue, not the whole solution. A couple of version notes: an early bug meant it couldn't read LoRAs when the loader's model was a qwen, flux, or lumina2 (Z-IMG) type, fixed in v1.3.1. And if the lookup fails, check the node's ID hasn't changed after a graph edit - the link input avoids that whole class of problem.