LoRA Params Loader v03

You've written a great LoRA library card with the Info Setter - trigger words, strengths, architecture - and now you're back in a fresh workflow. Do you retype all of it from memory? No. LoRA Params Loader v03 is the read side of that bookkeeping: give it a lora_path and it pulls the saved entry out of the pack's LoRA database and hands you the trigger words and strengths as clean outputs you can wire straight into a prompt builder or a LoRA loader. It's the node that closes the loop between "I annotated this" and "I actually use this."

The interface could not be simpler: one required input, lora_path, and nothing optional. The outputs are where the value lives: trigger_words_all and trigger_words_selected (both STRING - the full list and your chosen subset, ready to splice into a prompt), model_strength and clip_strength (both FLOAT - the strengths you recorded), and architecture and category (both STRING - for display or for conditional logic downstream).

The wiring that makes it shine: pipe trigger_words_selected into whatever builds your positive prompt, and pipe model_strength/clip_strength into the strength inputs of your LoRA loader node. If you've got a switch or condition, you can even branch on architecture - "if SDXL, load the SDXL variant; if Flux, load this other one." That's the whole game: the annotations you made once get reused everywhere, consistently, with no retyping drift.

Mechanically it's a database lookup keyed by the LoRA's path in the shared lora_tester_db.json. No model loading, no network - instant. The dependency chain is worth knowing: this node is only as good as the entries in the store. If you point it at a LoRA that was never annotated (or whose entry was cleaned out by the maintenance node), you'll get empty strings and defaults, which looks like a bug until you realize it's just an empty library card. Run the family's Info Setter first if the outputs come back blank.

Honest framing: this is a single-author, niche pack with no community footprint, and the Params Loader only makes sense if you're buying into the family's database idea. If you are, it's quietly one of the most convenient nodes in the group - it converts "notes you took once" into "parameters you never have to retype." If you're not annotating LoRAs, there's nothing here for you yet, and that's fine.

Install the pack and restart, then find it under loaders/lora tester:

cd ComfyUI/custom_nodes
git clone https://github.com/EricRollei/AAA_Metadata_System
cd AAA_Metadata_System
pip install -r requirements.txt

Set lora_path, run it, and wire the string and float outputs into your prompt and loader. That's the entire node - and that's the point.