ComfyUI-DvD-LoraTriggerwordsManager

ComfyUI-DvD-LoraTriggerwordsManager

中文说明

A set of professional ComfyUI custom nodes designed to manage LoRA trigger words efficiently. Seamlessly read, edit, and save trigger words in WebUI (Automatic1111) compatible JSON format.

✨ Features

• 📖 Auto-Read: Automatically reads .json or .txt metadata associated with your LoRA from the same directory.
• 🌐 Smart Lookup: (New!) If no local metadata exists, it automatically calculates the model hash (AutoV3/V2), queries Civitai for trigger words, and generates a config file.
• ✍️ Read & Edit: Not only reads tags but allows you to edit and save them directly within the node.
• 🔄 WebUI Alignment: Saves files strictly following the A1111 WebUI standard JSON format (e.g., "activation text", "sd version"), ensuring full cross-software compatibility and data preservation.
• 🔗 Auto-Merge: Whether chaining single nodes or using a stack, all trigger words are automatically concatenated with commas. No need for extra text concatenation nodes.
• 🐍 Pure Python: Zero frontend dependencies, stable performance, and no complex installation steps.

📦 Nodes

1. DvD LoRA Loader (Trigger Words)

Standard LoRA loader with MODEL and CLIP connections.

• Inputs: Model, Clip, LoRA Name, Strength.
• Mode: Read (Default) or Save.
• Edit Text: Input new trigger words here when in Save mode.
• Pre_text (Optional): Connect string from previous node to append.

2. DvD LoRA Loader (Model Only)

Optimized for FLUX/SD3 or workflows where you only need to modify the model weights without affecting CLIP directly (or handling CLIP separately).

3. DvD LoRA Stack (Multi-Merge)

Load 3 LoRAs at once.

• Merge Logic: Automatically combines trigger words from all 3 LoRAs + pre_text.
• Save Target: Select which LoRA (1, 2, or 3) to update when saving.

🚀 Installation

1. Navigate to your ComfyUI custom nodes directory:

   cd ComfyUI/custom_nodes/
   

2. Clone this repository:

   git clone https://github.com/idvdii/ComfyUI-DvD-LoraTriggerwordsManager.git
   

3. Restart ComfyUI.

🛠 Usage

1. Read Mode

Just select a LoRA. The node will automatically read the associated trigger words.

2. Save Mode

• Change mode to Save.
• Type your new tags in edit_text.
• Queue a prompt (run once).
• Result: The JSON file is updated/created instantly.

3. Stack Mode

Multiple LoRAs in one node. Trigger words are automatically concatenated. You can also specify which LoRA to update using save_target.

4. Compatibility

The generated JSON files use the standard format ("activation text", etc.), ensuring full compatibility with Stable Diffusion WebUI.

| WebUI Format vs Plugin Format | Smart Update (Preserve Data) | | :---: | :---: | | | |

5. Automatic Discovery

Scenario: You downloaded a "naked" LoRA (.safetensors only) and don't know the trigger words.

1. The initial folder (No JSON):

2. Automatic Online Lookup: Just run the node. It calculates the hash, queries Civitai, and finds the correct tags (e.g. "MMD, 3D").

3. Generation & Result: The tags are automatically injected into the prompt to generate the image.

4. Permanent Saving: A standard .json file is automatically created with the correct format. You never need to look it up again!

---

License: MIT