RandomLoRALoader

Random LoRA Loader for ComfyUI

English | 日本語版 README

A ComfyUI custom node package for randomly selecting and applying LoRAs. Includes three nodes:

1. Random LoRA Loader - Select from 3 folders simultaneously
2. Filtered Random LoRA Loader - Single folder with keyword filtering
3. Filtered Random LoRA Loader (LBW) - 🆕 LBW support with automatic SD1.5/SDXL detection (NEW v1.2.0)

Random LoRA Loader Workflow Example Filtered Random LoRA Loader Workflow Example FilteredRandomLoRALoader(LBW) Workflow Example

---

⚠️ Important: Supported Models

These nodes are designed for SD1.5 and SDXL models ONLY.

• ✅ Supported: Stable Diffusion 1.5, Stable Diffusion XL (SDXL)
• ❌ NOT Supported: Flux, SD3, SDXL Turbo, Pony, or other architectures

The LBW node's block weight feature specifically targets SD1.5/SDXL U-Net architecture. Other model types will not work correctly with LBW.

---

Nodes Overview

Random LoRA Loader (Original)

Select LoRAs from up to 3 different folders. Ideal for folder-based organization.

Use case:

• Different folders for styles, characters, concepts
• Fixed LoRA categories
• Simple folder-based workflow

Filtered Random LoRA Loader

Select LoRAs from a single folder using keyword filtering. Ideal for dynamic selection and large collections.

Use case:

• All LoRAs in one folder, filter by keywords
• Dynamic filtering with AND/OR conditions
• Metadata search capability
• Recommended for chaining multiple instances

Filtered Random LoRA Loader (LBW) 🆕 NEW v1.2.0

Advanced node with LoRA Block Weight (LBW) support for precise effect control.

Use case:

• Separate control of style vs. structure
• Fine-tuned LoRA application
• Professional workflows requiring precision
• 4 preset modes + custom input

See README_LBW.md for detailed LBW documentation.

---

Key Features

Common Features (All Nodes)

• Multi-Source Metadata Reading: Automatically retrieve trigger words and sample prompts with priority order:
  1. .metadata.json format (ComfyUI Lora Manager) - Priority 1
  2. .info format (Civitai Helper) - Priority 2
  3. Embedded metadata in LoRA file - Priority 3
• Strength Randomization:
  • Range specification with negative support: 0.4-0.8, -0.8--0.3 (v1.2.0)
  • 0.1 increments for ranges: Easier to observe effect differences
  • Fixed values rounded to 2 decimals: 0.847 → 0.85
• Flexible Trigger Word Retrieval:
  • json_combined: Combine all trigger word patterns (with deduplication)
  • json_random: Randomly select one pattern
  • json_sample_prompt: Randomly retrieve sample prompts
  • metadata: Read directly from embedded metadata
• Sample Prompt Optimization: Automatically remove LoRA syntax from samples and apply node-configured strength
• Text Outputs: positive_text / negative_text, plus lora_text carrying the LoRA syntax (v1.4.0)
• ComfyUI Standard Seed Control: Supports fixed/randomize/increment/decrement
• Wildcard Encode Integration: Works seamlessly with Wildcard Encode (Inspire)
• LoRA Syntax Auto-Removal: Cleans up LoRA syntax in additional_prompt to prevent noise
• Preview Image Support: Display preview images/videos for selected LoRAs (v1.1.0)
• Duplicate Filename Handling: Exclude duplicate filenames across subfolders (v1.1.0)

Filtered Nodes (Filtered & LBW)

• Keyword Filtering: Space-separated keywords with phrase support
• AND/OR/OFF Modes: Flexible filtering logic (OFF: v1.5.0)
• Metadata Search: Search in filenames or embedded metadata
• Fast Caching: Instant metadata access after first load
• Keyword Output: Outputs the keywords the selected LoRA actually matched (v1.5.0)

LBW Node Only 🆕

• LoRA Block Weight Support: Control which U-Net blocks are affected
• **Aut