Extensions/ComfyUI-LoraTags
ComfyUI Extension

ComfyUI-LoraTags

A powerful, sleek, and highly optimized LoRA loader for ComfyUI with modern tag management, automated Civitai metadata fetching, and seamless multi-LoRA stacking.

By iiTzMYUNG·Created about a month ago·Updated about a month ago· 6
iiTzMYUNG/ComfyUI-LoraTags
Nodes2
On cloudLocal install
Categoryconditioning, loaders
Stars6
Updatedabout a month ago
Readme

ComfyUI-LoraTags 🏷️

ComfyUI License

ComfyUI-LoraTags is a sleek, highly optimized LoRA loader for ComfyUI. It replaces the standard LoRA loading experience with a modern tag management system, automated Civitai metadata fetching, and seamless multi-LoRA stacking — all inside a single, thoughtfully designed custom node.

ComfyUI LoraTags Node & Auto-Tags Text Encoder


✨ Key Features

📚 Infinite LoRA Stacking

Keep your workflows clean. Instead of daisy-chaining multiple standard LoRA loaders, LoraTags lets you stack as many LoRAs as you need inside a single node. Toggle them on/off, adjust strengths, or delete them with a single click.

🌐 Automated Civitai Integration

Click the i (Info) button on any LoRA to open the Tag Manager modal. The plugin calculates the SHA256 hash of your .safetensors file and queries the Civitai API to fetch:

  • The official model name
  • The original trained trigger words
  • Up to 4 preview images
  • A direct link to the model's Civitai page

🧠 Master Database Tag Manager

No more .txt companion files cluttering your hard drive. All tags are saved to a centralized lora_master_tags.json file inside the plugin folder.

  • Locked tags (green bubbles): retrieved automatically from Civitai
  • Custom tags (blue bubbles): typed in manually — add, edit, or remove at will

🔗 Auto-Tagging CLIP Text Encoder

Route all your active LoRA trigger words directly into your prompt. Connect the TAGS output to the custom CLIP Text Encode (Auto-Tags) node, and your saved triggers are automatically appended to your prompt. Active tags are displayed as responsive bubbles right on the canvas.

📋 One-Click Clipboard Copying

Prefer manual prompting? Your saved trigger words are shown right on the canvas beneath each LoRA. Click the 📋 icon to instantly copy them to your clipboard.

🎨 Modern Canvas UI

Built with a custom HTML5 Canvas rendering engine that bypasses the default LiteGraph widgets, featuring:

  • Smooth, iOS-style toggle switches
  • Glassmorphism panels with semi-transparent borders
  • Clean, minimalist typography and icons

📦 Installation

Method 1 — ComfyUI Manager (coming soon)

Once approved in the ComfyUI Manager registry, you'll be able to search for ComfyUI-LoraTags in the custom nodes list and install with one click.

Method 2 — Git Clone (recommended)

cd ComfyUI/custom_nodes/
git clone https://github.com/iiTzMYUNG/ComfyUI-LoraTags.git

Then restart ComfyUI and hard-refresh your browser (Ctrl+Shift+R).

Method 3 — Manual Download

  1. Click the green <> Code button at the top of this repo and select Download ZIP.
  2. Extract the .zip file.
  3. Move the extracted ComfyUI-LoraTags folder into your ComfyUI/custom_nodes/ directory.
  4. Restart ComfyUI and hard-refresh your browser.

🚀 How to Use

  1. Add the nodes — double-click the canvas and search for LoRA Loader (LoraTags) and CLIP Text Encode (Auto-Tags).
  2. Connect the wires — link the TAGS output from the LoRA Loader to the tags input of the Auto-Tags Text Encoder.
  3. Stack LoRAs — click + Add LoRA to create a new slot, then choose a .safetensors file from the dropdown.
  4. Manage tags — click the i icon on any LoRA row to fetch tags from Civitai or add your own.
  5. Generate — type your base prompt; active tags are appended automatically.

📂 File Structure & the Master Database

Saving tags generates a lora_master_tags.json file inside the ComfyUI-LoraTags directory.

The plugin respects your existing folder structure — for example, models/loras/SDXL/neon.safetensors will be categorized under an SDXL header in the JSON. This keeps the database portable, so you can back it up or share it with friends as a single file.


🤝 Contributing

Contributions, issues, and feature requests are welcome! Check the issues page if you'd like to help improve the UI or add functionality.

📜 License

This project is licensed under the MIT License — see the LICENSE file for details.