🎨 ComfyUI Prompt Expander

An LLM-powered prompt expansion node for ComfyUI. Transform simple prompts into detailed, high-quality image generation prompts.

✨ Features

• 🚀 Groq API support (free, super fast)
• 🏠 Ollama support (local, offline)
• 🎬 5 Styles: cinematic, anime, photorealistic, artistic, niji
• 📊 4 Detail Levels: minimal, medium, detailed, extreme
• ✅ Automatic positive & negative prompt generation
• 🔧 Custom instructions support

📦 Installation

Method 1: ComfyUI Manager (Recommended)

Search for "Prompt Expander" in ComfyUI Manager and install.

Method 2: Manual Installation

cd ComfyUI/custom_nodes
git clone https://github.com/bozkut/ComfyUI-Prompt-Expander.git

Restart ComfyUI after installation.

🔑 API Setup

Groq (Recommended - Free & Fast)

1. Go to console.groq.com
2. Create a free account
3. Navigate to API Keys → Create API Key
4. Copy and paste the key into the node, or set as environment variable:

Windows (Command Prompt):

setx GROQ_API_KEY "gsk_xxxxxxxxxxxxx"

Windows (PowerShell):

[Environment]::SetEnvironmentVariable("GROQ_API_KEY", "gsk_xxxxxxxxxxxxx", "User")

Linux/Mac:

export GROQ_API_KEY="gsk_xxxxxxxxxxxxx"

Ollama (Local/Offline)

# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh

# Download a model
ollama pull llama3.2

# Start Ollama server
ollama serve

🎯 Usage

Node: 🎨 Prompt Expander (LLM)

Located in: bozkut/AI

Input	Description
simple_prompt	Your basic prompt (e.g., "boy, blonde, driving a car")
llm_provider	Choose between groq or ollama
model	Select LLM model
style	cinematic, anime, photorealistic, artistic, niji
detail_level	minimal, medium, detailed, extreme
groq_api_key	Your Groq API key (optional if using env variable)
custom_instructions	Additional instructions for the LLM (optional)
negative_prompt_request	Generate negative prompt (true/false)

Output	Description
positive_prompt	Expanded detailed prompt
negative_prompt	Generated negative prompt

📸 Example

Input:

boy, blonde, driving a car in San Francisco

Output (cinematic, detailed):

A young blonde-haired man behind the wheel of a classic convertible, 
golden hour sunlight streaming through the windshield, the iconic 
Golden Gate Bridge visible in the distant background, warm amber and 
teal color grading, shallow depth of field with bokeh lights, 
cinematic composition, lens flare, 8k resolution, photorealistic 
skin texture, wind-swept hair, California coastal atmosphere, 
dramatic lighting, movie still quality

🔗 Workflow

Connection Diagram

┌─────────────────────┐
│  🎨 Prompt Expander │
│  "simple prompt"    │
└──┬──────────────┬───┘
   │              │
   ▼              ▼
┌────────┐    ┌────────┐
│positive│    │negative│
│ prompt │    │ prompt │
└───┬────┘    └───┬────┘
    │             │
    ▼             ▼
┌────────┐    ┌────────┐
│ShowText│    │ShowText│  (optional - to preview)
└───┬────┘    └────────┘
    │
    ▼
┌─────────────────────┐
│  Image Generator    │
│  (Z Image, SDXL..)  │
└─────────────────────┘

Included Workflows

• workflow_z_image_turbo.json - Ready-to-use workflow with Z Image Turbo

🛠️ Requirements

• ComfyUI
• Python 3.10+
• requests library (usually pre-installed)

Optional

• ComfyUI-Custom-Scripts - For ShowText node

🤝 Contributing

Pull requests are welcome! Feel free to:

• Report bugs
• Suggest new features
• Add new styles or LLM providers

📄 License

MIT License - see LICENSE for details.

🙏 Credits

Created by Bora Ozkut

---

⭐ If you find this useful, please consider giving it a star!