Chye ComfyUI Toolset
A comprehensive collection of ComfyUI custom nodes organized by category for enhanced workflow productivity.
🎯 Features
🔹 Latent Tools
- Model-Specific Aspect Ratios: Optimized presets for Flux, Qwen Image, and SDXL models
- Phone Screen Sizes: Common mobile device resolutions including modern tall screens
- Video Formats: Standard video resolutions from 480p to 4K including cinema formats
- Social Media Formats: Optimized sizes for Instagram, TikTok, YouTube, Facebook, and Twitter
- Smart Orientation Control: Automatic Portrait/Landscape switching
- Precision Scaling: Multiplier with automatic rounding to multiples of 32
- Batch Generation: Support for multiple latents at once
🔸 Post Process Tools
- Realistic Film Grain: Apply photographic film grain simulation with ISO control
- ARRI Halation Effect: Simulate film bloom around highlights with red-orange tint
- Chromatic Aberration: Lens color fringing simulation with barrel distortion
- Spatially Correlated Noise: Advanced noise generation using scipy for realistic grain patterns
- Color Channel Control: Option for monochrome or colored grain effects
🔤 Prompt Tools
- Multi-Provider LLM Integration: DeepSeek, OpenAI, Anthropic, and custom providers
- Dual Template Systems: File-based templates AND editable text input options
- Secure API Key Management: File-based storage with gitignore protection
- Automatic Setup: Dedicated setup node creates folder structure with examples
- Professional Templates: Photography, cinematic, and high-detail enhancement templates
- Error Handling: Graceful degradation with comprehensive error reporting
🔧 Utility Tools (Coming Soon)
- Workflow helpers
- Data conversion utilities
- Debug and analysis tools
📦 Installation
Method 1: Git Clone (Recommended)
cd custom_nodes
git clone https://github.com/chyer/Chye-ComfyUI-Toolset.git
cd Chye-ComfyUI-Toolset
echo "Chye-ComfyUI-Toolset" > .git/.cnr-id
# Install dependencies
pip install -r requirements.txt
Method 2: Manual Copy
- Copy the
Chye-ComfyUI-Toolsetfolder to your ComfyUIcustom_nodesdirectory - Important: Create a file named
.cnr-idin the.gitfolder with contentChye-ComfyUI-Toolsetto avoid workflow errorsecho "Chye-ComfyUI-Toolset" > .git/.cnr-id - Install dependencies:
cd Chye-ComfyUI-Toolset pip install -r requirements.txt - Restart ComfyUI
🚀 Usage
Latent Tools
Find the nodes under the latent category:
Model-Specific Nodes
- 🔹 CYH Latent | Flux Aspect Ratio - Flux-optimized latent generation
- 🔹 CYH Latent | Qwen Aspect Ratio - Qwen Image-optimized latent generation
- 🔹 CYH Latent | SDXL Aspect Ratio - SDXL-optimized latent generation
Phone Screen Nodes
- 🔹 CYH Latent | Phone Aspect Ratio - Common mobile device resolutions
Video Format Nodes
- 🔹 CYH Latent | Video Aspect Ratio - Standard video resolutions and formats
Social Media Nodes
- 🔹 CYH Latent | Social Media Aspect Ratio - Optimized sizes for social platforms
Configuration Options:
- Aspect Ratio: Choose from category-specific presets
- Orientation: Portrait or Landscape (with smart defaults based on media type)
- Multiplier: Scale resolution (0.1-10.0)
- Batch Size: Number of latents to generate (1-64)
Example Workflow:
[🔹 CYH Latent | Flux Aspect Ratio] → [KSampler] → [VAE Decode]
Multiplier Example:
Model: SDXL
Aspect: 16:9
Orientation: Landscape
Multiplier: 1.5
→ Output: 1536×864 latent
🎬 Post Process Tools
Find the nodes under the post_process category:
-
🎬 CYH Post Process | Film Grain - Apply realistic film grain simulation
- Strength: Overall grain intensity (0.0 to 1.0)
- ISO: Simulated film sensitivity (100-6400)
- Grain Size: Particle size control (1.0-10.0)
- Colored: Toggle between monochrome or colored grain
-
🎬 CYH Post Process | ARRI Halation - Apply ARRI-style film bloom effect
- Threshold: Highlight detection threshold (0-255)
- Blur Size: Glow effect size (1-101, odd numbers)
- Intensity: Halation strength (0.0 to 1.0)
-
🌈 CYH Post Process | Chromatic Aberration - Apply lens color fringing effect
- Preset: Quick configurations (None, Vintage, Modern, Extreme, Custom)
- Intensity: Master effect strength multiplier (0.0-2.0)
- Center: Adjustable distortion center point (0.0-1.0)
- Quality: Fast (bilinear) or High (bicubic) interpolation
- Advanced Controls: Individual RGB channel distortion coefficients (k1, k2, k3)
Preset Details:
- None: No chromatic aberration
- Vintage: Classic lens look with red/blue separation
- Modern: Subtle modern lens character
- Extreme: Strong artistic effect with complex distortion
Example Workflows:
[KSampler] → [VAE Decode] → [🎬 CYH Post Process | Film Grain]
[KSampler] → [VAE Decode] → [🎬 CYH Post Process | ARRI Halation]
[KSampler] → [VAE Decode] → [🌈 CYH Post Process | Chromatic Aberration]
## 🔤 Prompt Tools
Find the nodes under the **prompt_tools** category:
- **🔤 CYH Prompt Tools | Setup** - Creates folder structure and example templates
- **Force Recreate**: Overwrite existing files (True/False)
- **Output**: Status message with setup results
- **🔤 CYH Prompt Tools | Enhancer** - Uses LLM APIs to refine and enhance prompts
- **Prompt**: Input prompt to enhance (multiline text)
- **Template Path**: Path to template file (e.g., "providers/deepseek/professional_photo.txt")
- **Provider**: LLM provider (DeepSeek, OpenAI, Anthropic, Custom)
- **API Key Override**: Optional API key override (skips file lookup)
- **Temperature**: Creativity control (0.0-1.0)
- **Max Tokens**: Response length limit (100-4000)
- **🔤 CYH Prompt Tools | Enhancer (Editable)** - Alternative version with editable preprompt template
- **Prompt**: Input prompt to enhance (multiline text)
- **Preprompt Template**: Editable template text (must include {prompt} placeholder)
- **Provider**: LLM provider (DeepSeek, OpenAI, Anthropic, Custom)
- **API Key Override**: Optional API key override (skips file lookup)
- **Temperature**: Creativity control (0.0-1.0)
- **Max Tokens**: Response length limit (100-4000)
### Setup Instructions:
1. **Run Setup Node First**: Use "🔤 CYH Prompt Tools | Setup" to create folder structure
2. **Configure API Keys**: Edit files in `preprompts/api_keys/` with your actual API keys
3. **Customize Templates**: Modify or create templates in `preprompts/` subdirectories
4. **Use Enhancer**: Connect prompts to "🔤 CYH Prompt Tools | Enhancer" for refinement
### Example Templates:
- **Professional Photography**: Camera settings, lighting, composition techniques
- **Cinematic Style**: Film terminology, director styles, visual storytelling
- **High Detail**: Ultra-realistic textures, technical precision, intricate patterns
### Supported Providers:
- **DeepSeek**: Fast and cost-effective Chinese LLM
- **OpenAI**: GPT-4 Turbo for high-quality enhancements
- **Anthropic**: Claude models for nuanced prompt refinement
- **Custom**: Template-only mode without API calls
### Example Workflows:
[Text Input] → [🔤 CYH Prompt Tools | Enhancer] → [KSampler]
[🔤 CYH Prompt Tools | Setup] → [Configure API Keys] → [🔤 CYH Prompt Tools | Enhancer]
[Text Input] → [🔤 CYH Prompt Tools | Enhancer (Editable)] → [KSampler]
### When to Use Each Version:
- **File-based Enhancer**: Use when you want to maintain reusable templates across workflows
- **Editable Enhancer**: Use when you want to quickly modify or experiment with preprompt text directly in the node
- **Both versions support the same API providers and configuration options**
### Folder Structure:
preprompts/ ├── api_keys/ # API key files (gitignored) │ ├── deepseek_api_key.txt │ ├── openai_api_key.txt │ └── anthropic_api_key.txt ├── providers/ # Provider-specific templates │ ├── deepseek/ │ ├── openai/ │ └── anthropic/ ├── styles/ # Artistic style templates ├── quality/ # Quality enhancement templates ├── models/ # Model-specific optimizations └── custom/ # User-defined templates
📱 Phone Screen Resolutions
| Aspect Ratio | Resolution | Description |
|---|---|---|
| 16:9 | 1080×1920 | Standard phone screens |
| 19.5:9 | 1080×2340 | Modern tall phones |
| 20:9 | 1080×2400 | Ultra tall phones |
| 18:9 | 1080×2160 | Tall phones |
Default Orientation: Portrait (optimized for mobile viewing)
🎬 Video Format Resolutions
| Aspect Ratio | Resolution | Description |
|---|---|---|
| 16:9 | 1920×1080 | Standard video (Full HD) |
| 21:9 | 1920×810 | Cinema format |
| 4:3 | 1024×768 | Traditional video |
| 9:16 | 1080×1920 | Portrait video |
Default Orientation: Landscape (except 9:16 which defaults to Portrait)
📱 Social Media Resolutions
| Platform | Format | Resolution | Aspect Ratio |
|---|---|---|---|
| Square | 1080×1080 | 1:1 | |
| Portrait | 1080×1350 | 4:5 | |
| Landscape | 1080×608 | 1.91:1 | |
| Stories/Reels | 1080×1920 | 9:16 | |
| TikTok | Standard | 1080×1920 | 9:16 |
| TikTok | Wide | 1920×1080 | 16:9 |
| YouTube | Standard | 1920×1080 | 16:9 |
| YouTube | Shorts | 1080×1920 | 9:16 |
| Feed | 1200×630 | 1.91:1 | |
| Stories | 1080×1920 | 9:16 | |
| Tweet | 1200×675 | 16:9 | |
| Header | 1500×500 | 3:1 |
Default Orientation: Platform-specific (Portrait for Instagram Stories/TikTok, Landscape for YouTube Standard, etc.)
📐 Resolution Reference
| Model | 1:1 | 4:3 | 3:2 | 16:9 | 21:9 |
|---|---|---|---|---|---|
| Flux | 1024×1024 | 1280×960 | 1152×768 | 1344×768 | 1792×768 |
| Qwen | 1328×1328 | 1472×1140 | 1536×1024 | 1664×928 | 1984×864 |
| SDXL | 1024×1024 | 1024×768 | 1152×768 | 1024×576 | 1344×576 |
🏗️ Project Structure
Chye-ComfyUI-Toolset/
├── categories/
│ ├── latent_tools.py # Aspect ratio latent generators
│ ├── post_process.py # Image post-processing tools (film grain, etc.)
│ ├── prompt_tools.py # LLM-based prompt enhancement tools
│ ├── image_tools.py # (Future) Image processing tools
│ ├── utility_tools.py # (Future) Workflow utilities
│ └── __init__.py
├── shared/
│ ├── constants.py # Common configuration data
│ ├── validators.py # Input validation utilities
│ ├── helpers.py # Shared utility functions
│ └── __init__.py
├── preprompts/ # Prompt enhancement templates and API keys
│ ├── api_keys/ # API key files (gitignored)
│ ├── providers/ # Provider-specific templates
│ ├── styles/ # Artistic style templates
│ ├── quality/ # Quality enhancement templates
│ ├── models/ # Model-specific optimizations
│ └── custom/ # User-defined templates
├── __init__.py # Main entry point
├── pyproject.toml # Project metadata
└── README.md # This file
📝 Notes
- All dimensions are automatically rounded to multiples of 32 for ComfyUI compatibility
- Portrait orientation swaps width/height when the base resolution is landscape-oriented
- Multiplier scales base resolution before rounding
- Modular design allows for easy expansion with new tool categories
- Phone screen nodes default to Portrait orientation for optimal mobile viewing
- Video format nodes have smart defaults based on the aspect ratio (Landscape for most, Portrait for 9:16)
- Social media nodes have platform-specific default orientations for optimal results
🔄 Version History
- v2.4.0: Added PromptEnhancerEditable node with editable preprompt templates
- v2.3.0: Added Prompt Enhancement tools with multi-provider LLM integration
- v2.2.0: Added Chromatic Aberration node with realistic lens simulation
- v2.1.0: Added Phone, Video, and Social Media aspect ratio nodes
- v2.0.0: Complete restructure into modular toolset with shared utilities
- v1.x.x: Individual aspect ratio nodes
🤝 Contributing
This toolset is designed for extensibility. New categories and tools can be easily added following the established patterns in the categories/ and shared/ directories.
Happy Creating! 🎨