light-heart-labs--dreamserver/ods/extensions/library/workflows/comfyui

ComfyUI Workflows

This directory contains n8n workflow templates for integrating ComfyUI with other services.

Available Workflows

1. LLM to Image Generation (llm-to-image.json)

This workflow enhances prompts with the LLM, then uses a predefined ComfyUI workflow with the enhanced prompt:

1. Webhook: Receives prompts with optional API key authentication
2. Input Validation: Validates prompt length, channel format
3. LLM Enhancement: Sends to LLM to generate an enhanced image generation prompt
4. ComfyUI Execution: Uses the predefined workflow with the enhanced prompt
5. Output: Returns enhanced prompt and status to Discord

Use Case

Best for consistent, high-quality image generation with prompt engineering.

Environment Variables

• LLM_HOST - LLM server host (default: localhost)
• LLM_PORT - LLM port (default: 8080)
• COMFYUI_HOST - ComfyUI host (default: localhost)
• COMFYUI_PORT - ComfyUI port (default: 8188)
• COMFYUI_MODEL_NAME - ComfyUI model name (default: flux1-dev.safetensors)
• LLM_TO_IMAGE_API_KEY - Optional API key for webhook authentication

Usage

curl -X POST http://localhost:5678/webhook/llm-to-image \
  -H "Content-Type: application/json" \
  -H "X-API-Key: your-api-key" \
  -d '{
    "prompt": "A beautiful landscape with mountains",
    "channel": "#art"
  }'

2. Dynamic Workflow Generation (llm-image-gen.json)

This workflow enables AI-powered image generation by having the LLM generate the ComfyUI workflow JSON dynamically:

1. Webhook: Receives image generation prompts via POST to /comfyui-dynamic-workflow
2. LLM Processing: Sends the prompt to the LLM (llama-server) to generate a ComfyUI workflow JSON
3. ComfyUI Execution: Submits the dynamically generated workflow to ComfyUI for image generation
4. Output: Returns status to Discord

Use Case

Best when you want the LLM to adapt the ComfyUI workflow structure based on the prompt complexity and requirements.

Environment Variables

• LLM_HOST - LLM server host (default: localhost)
• LLM_API_PORT - LLM API port (default: 8080)
• COMFYUI_HOST - ComfyUI host (default: localhost)
• COMFYUI_PORT - ComfyUI port (default: 8188)
• COMFYUI_DYNAMIC_API_KEY - Optional API key for webhook authentication

Usage

curl -X POST http://localhost:5678/webhook/comfyui-dynamic-workflow \
  -H "Content-Type: application/json" \
  -H "X-API-Key: your-api-key" \
  -d '{
    "prompt": "Create a photorealistic portrait of a cyberpunk samurai",
    "channel": "#art"
  }'

Setup

1. Import workflows into n8n
2. Configure environment variables in your n8n instance
3. Ensure ComfyUI is running and accessible
4. Ensure LLM server (llama-server) is running and accessible