⚠️ IMPORTANT: Installation for Audio Enhancement Features

If you want the voice enhancement and advanced audio processing nodes (Seed-VC integration), you MUST complete the additional setup:

Option 1: Portable ComfyUI (Recommended)

Simply run the provided batch file:

Portable_auto_install_nodes.bat

This will automatically:

• Initialize the seed-vc git submodule
• Install all required dependencies
• Set up audio processing packages

Option 2: Manual Installation (venv/conda)

If you're using a virtual environment or conda:

Step 1: Initialize seed-vc submodule

cd /path/to/ComfyUI/custom_nodes/Boyonodes
git submodule add https://github.com/Plachtaa/seed-vc.git seed-vc
git submodule update --init --recursive

Step 2: Install dependencies

# Core audio dependencies
pip install librosa>=0.10.0 transformers>=4.30.0 safetensors>=0.3.0
pip install huggingface_hub>=0.15.0 pyloudnorm>=0.1.0 soundfile>=0.12.0

# Seed-VC specific requirements
pip install hydra-core>=1.3.0 omegaconf munch descript-audio-codec

# TTS packages (no dependencies to avoid conflicts)
pip install chatterbox-tts --no-deps
pip install resemble-perth --no-deps

# Install remaining requirements
pip install -r requirements.txt

Step 3: Restart ComfyUI

---

Boyonodes

Essential ComfyUI nodes for semantic image editing, audio processing, LoRA management, and automated workflow generation. Streamlines complex pipelines with intelligent automation and robust error handling.

🚀 Quick Installation

git clone https://github.com/DragonDiffusionbyBoyo/Boyonodes.git
cp -r Boyonodes /path/to/ComfyUI/custom_nodes/

Restart ComfyUI after installation. Most nodes work immediately with no additional dependencies.

📋 Installation Requirements

Core Nodes (No additional dependencies)

• Semantic Image Editing nodes
• LoRA Management System
• Workflow Enhancement nodes
• Basic Utility nodes

Audio Processing Nodes (Basic)

pip install librosa transformers safetensors huggingface_hub pyloudnorm soundfile
pip install chatterbox-tts --no-deps
pip install resemble-perth --no-deps

Advanced Audio Enhancement (Seed-VC)

See installation instructions at the top of this README

Mandelbrot Video Generator

pip install numpy==1.26 matplotlib pillow tqdm torch

FFmpeg (for video output)

• Windows: Download from ffmpeg.org or choco install ffmpeg
• macOS: brew install ffmpeg
• Linux: sudo apt-get install ffmpeg

---

🎵 Audio Processing & TTS

Chatterbox Turbo TTS Integration

High-quality voice cloning and text-to-speech generation with emotion support.

Key Features:

• Voice Cloning: Clone any voice from 5+ second reference audio
• Emotion Tags: Natural expressions like [chuckle], [sigh], [gasp]
• GPU Accelerated: CUDA support for fast generation
• 24kHz Quality: Professional audio output

Available Emotion Tags: [clear throat] [sigh] [shush] [cough] [groan] [sniff] [gasp] [chuckle] [laugh]

Workflow:

1. Boyo Chatterbox Turbo Loader - Load TTS model
2. Boyo Chatterbox Turbo Generate - Create speech with emotion
3. Standard ComfyUI audio nodes for processing

Audio Padding & Synchronization

Finally solved the community's lip-sync timing challenge! Precision audio padding for perfect video-audio synchronization.

Key Nodes:

• BoyoAudioDurationAnalyzer - Extract precise duration from any audio tensor
• BoyoAudioPadder - Intelligent silence padding with auto-centering

Why These Nodes Succeed:

• ✅ Handles all ComfyUI audio formats ((batch,channels,samples) vs (batch,samples,channels))
• ✅ Automatic 12kHz output for proper playback speed
• ✅ Intelligent feedback: 📊 Audio: 3.2s + Padding: 7.3s = Total: 10.5s ✅ Perfect match!
• ✅ Memory-efficient processing for large files
• ✅ Auto-centering with target duration matching

Quick Workflow:

Load Audio → BoyoAudioPadder (target_duration: 10.5s, auto_center: true) → Save Audio

---

🎨 Semantic Image Editing System

Perfect for Kontext, Qwen Image Edit, and HiDream E1.1 workflows with automated iteration and dataset creation.

Core Editing Nodes

Boyo Image Grab

• Auto-monitors directories for newest images
• Enables seamless iterative editing chains
• Perfect for progressive semantic modifications
• Real-time directory monitoring

Boyo Paired Image Saver

• Saves original/edited pairs with sequential naming
• Dataset creation for training workflows
• ControlNet format compatibility
• Organized file management

Boyo Incontext Saver

• Specialized for semantic editing outputs
• Dataset-ready organization
• Maintains editing relationships

Boyo Universal Image + Prompt Saver

• Strips metadata for clean publication
• Saves actual prompts (crucial for wildcard workflows)
• Creates organized image/text pairs
• Publication-ready outputs

Workflow Example

Original → Semantic Edit → Paired Saver → Image Grab (auto-feeds next iteration)

---

🎯 LoRA Management System

Revolutionary paired LoRA management with intelligent prompt handling for complex workflows requiring multiple LoRA types.

Boyo LoRA JSON Builder

Create and save LoRA configurations with flexible prompt management.

Key Features:

• Supports paired LoRAs (high/low noise variants)
• Multiple prompts per configuration
• Handles utility LoRAs (no prompts needed)
• Auto-saves to organized directory

Boyo LoRA Paired Loader

Load multiple LoRA configurations simultaneously with advanced prompt strategies.

Key Features:

• 3 simultaneous config slots for layered effects
• Prompt strategies: Mute, Concatenate, Merge per config
• Prompt modes: First Only, Cycle Through, Random (seed-based)
• 6 LoRA path outputs + 4 prompt string outputs
• Direct connection to standard LoRA loaders

Boyo LoRA Config Inspector

Preview and analyze LoRA configurations before loading.

Sample Output:

📋 LoRA Configuration: Character_Cyborg
🎯 LoRA FILES:
  📈 High Noise: ✅ cyborg_char_v2.safetensors
  📉 Low Noise: ✅ cyborg_char_v2_low.safetensors
  🎭 Type: PAIRED LoRA

💬 PROMPTS (3 total):
  1. cyborg woman, metallic skin, glowing eyes
  2. android female, chrome details, futuristic
  3. robotic humanoid, synthetic appearance

💡 USAGE RECOMMENDATIONS:
  • Use 'Cycle Through' for variety
  • Use 'Random' for experimentation

---

📖 AI Storyboard Generation

Automated storyboard creation using local ollama models for consistent multi-scene video workflows.

Boyo Storyboard Prompt

Intelligent prompt generator for structured storyboard sequences.

Key Features:

• Model-agnostic trigger words - works with any LoRA/video model
• Two modes: 6-scene storyboards or traveling prompt sequences
• Consistent character/style across all scenes
• Optimized for abliterated coder models (Qwen 30B A3B Coder recommended)

System Prompt 1 (Standard): 6 image + 6 video prompts for Next Scene LoRA workflows
System Prompt 2 (Traveling): 6 images + 6 multi-line video sequences for extended content

Boyo Storyboard Output

Parses ollama JSON responses into 12 separate prompt outputs for direct workflow integration.

Workflow:

Storyboard Prompt → ollama Generate → Storyboard Output → 12 individual prompts

---

🛠️ Utility & Enhancement Nodes

Asset Downloader System

One-click workflow dependency installation. Drop JSON manifest files to automatically download custom nodes, models, and dependencies.

Features:

• Automatic GitHub repository cloning
• Python dependency installation via pip
• Model downloads from direct URLs
• Available in Simple (auto-detect) and Advanced (custom paths) versions

Workflow Enhancement

• Boyo Empty Latent - Smart aspect ratio calculator
• Load Image List - Batch image processor for mass operations
• Boyo VAE Decode - Stealth NSFW filtering for controlled environments
• Boyo Tiled VAE Decode - Memory-efficient large image processing

Creative Tools

• Mandelbrot Video Generator - Fractal art for creative projects
• BoyoVision Node - Qwen2.5VL vision with abliterated model compatibility

---

📚 Workflow Examples

Multi-LoRA Character Generation

1. Create configs for utility (lightning), character, and style LoRAs
2. Load all three simultaneously in Paired Loader
3. Set strategies: utility = "Mute", character/style = "Concatenate"
4. Get combined prompts and all LoRA paths in one node

Iterative Semantic Editing

1. Load initial image
2. Apply semantic edit (Kontext/Qwen/HiDream)
3. Boyo Paired Image Saver stores original + edit
4. Boyo Image Grab auto-feeds edit for next iteration
5. Repeat for progressive modifications

Video Lip-Sync Workflow

1. Load Video → VideoHelperSuite Info → get duration
2. Generate TTS audio → BoyoAudioPadder with auto-center
3. Perfect timing match for lip-sync models

Storyboard-to-Video Pipeline

1. Configure story/character in Storyboard Prompt
2. Generate via ollama → parse with Storyboard Output
3. Connect 6 image outputs to Next Scene LoRA
4. Connect 6 video outputs to video generation

---

🔧 Troubleshooting

Audio Issues

• Audio too fast: Node automatically outputs 12kHz for proper ComfyUI playback
• TTS loading errors: Ensure dependencies installed with --no-deps flags
• CUDA problems: Set device to "cpu" in loader node

LoRA Management

• Config not loading: Use Inspector node to verify file paths and JSON syntax
• Missing LoRA files: Check paths use forward slashes, verify file existence
• Prompt issues: Inspector shows available prompts and recommendations

Semantic Editing

• Image Grab not updating: Verify directory path exists, check auto_refresh enabled
• Paired Saver failing: Confirm output directory exists and has write permissions
• Slow performance: Organize files into smaller subdirectories

Storyboard Generation

• Poor outputs: Use recommended ollama models (Qwen 30B A3B Coder Abliterated)
• JSON parse errors: Add verbosity instructions in additional_details field
• Avoid: Google models (Gemma), Meta coding variants, thinking models

---

🏗️ Node Categories

• Boyo/Audio/TTS - Text-to-speech and voice cloning
• Boyo/Audio/Analysis - Audio duration and analysis tools
• Boyo/Audio/Processing - Padding and timing control
• Boyo/Image/Editing - Semantic editing workflow tools
• Boyo/LoRA - LoRA management and configuration
• Boyo/Storyboard - AI storyboard generation
• Boyo/Utility - General workflow enhancement tools

---

🤝 Contributing

1. Fork the repository
2. Create feature branch (git checkout -b feature-name)
3. Commit changes
4. Push to branch
5. Open pull request

Documentation for new features is appreciated.

---

📄 License

MIT License - see LICENSE file for details.

Built by DragonDiffusionbyBoyo for the semantic editing revolution.

---

Note: Vision nodes currently disabled due to dependency conflicts. Resolution in progress.