ComfyUI by IO-AtelierTech

One-command AI workflow automation for creative professionals

No configuration. No GPU required. Just just start and create.

What You Get

• ✅ ComfyUI pre-configured with curated plugins
• ✅ Cloud GPU inference via fal.ai (no local GPU needed)
• ✅ AI workflow assistant via Claude Code MCP
• ✅ Video generation Luma, Kling, FLUX models ready to use
• ✅ Project-scoped all data stays in your project folder

Quick Start

# 1. Install just (one-time setup)
curl --proto '=https' --tlsv1.2 -sSf https://just.systems/install.sh | bash

# 2. Clone this template
git clone https://github.com/io-ateliertech/comfyui-template.git my-project
cd my-project

# 3. Add your fal.ai API key
cp .env.example .env
nano .env  # Add your FAL_KEY from https://fal.ai/dashboard/keys

# 4. Start ComfyUI
just start

Visit http://localhost:8188

Commands

just start    # Start ComfyUI (pulls Docker image if needed)
just stop     # Stop ComfyUI
just logs     # View logs
just restart  # Restart ComfyUI
just update   # Update to latest version
just status   # Check status

Directory Structure

my-project/
├── data/
│   ├── input/         # Put your input images here
│   ├── output/        # Generated content saved here
│   └── models/        # Downloaded models cached here
├── workflows-api/     # API format workflows (for MCP/automation)
├── workflows-ui/      # UI format workflows (load in ComfyUI editor)
├── .env               # Your API keys (gitignored)
└── docker-compose.yml # Docker configuration

For Creatives: What This Does

No GPU Needed

All AI processing runs on fal.ai cloud GPUs. Your computer just runs the interface.

Pre-configured Plugins

• FLUX: High-quality image generation
• Luma: Image-to-video with camera control
• FFmpeg: Video processing
• Text overlays: Add text to images

AI Assistant Integration

Install the MCP server to let Claude Code create workflows for you:

just install-mcp

Then copy .mcp.json to your Claude Code project config.

Workflows

Saving Workflows

In ComfyUI:

1. Create your workflow
2. Settings → Export (API Format) → Save to workflows-api/
3. Settings → Save → Save to workflows-ui/

Using Workflows

• API format: Use with Claude Code MCP for automation
• UI format: Load in ComfyUI editor for visual editing

Examples

Generate an Image

1. Open http://localhost:8188
2. Find "FluxDev_fal" node (Right-click → Add Node → image/fal)
3. Add prompt: "a futuristic city at sunset"
4. Click "Queue Prompt"
5. Find output in data/output/

Image to Video

1. Put image in data/input/my-image.jpg
2. Load image node
3. Connect to "LumaImageToVideo_fal"
4. Add camera motion with "LumaConceptsNode"
5. Generate video → saved to data/output/video/

Requirements

• Docker - Install Docker
• just - Install just
• fal.ai API key - Get key

Updating

just update

Pulls the latest Docker image with updated plugins and models.

Troubleshooting

ComfyUI won't start

just logs  # Check for errors

Common issues:

• .env file missing or invalid FAL_KEY
• Port 8188 already in use

Models not loading

The first time you use a model, it downloads to data/models/. This can take time depending on model size.

Out of memory errors

Close other applications or increase Docker's memory limit:

• Docker Desktop → Settings → Resources → Memory

What's Inside the Docker Image?

The ioateliertech/comfyui-bundle image includes:

• ComfyUI (latest stable)
• GenAI Connectors (fal.ai, Replicate support planned)
• Video Utils (FFmpeg integration)
• Community plugins (FFmpeg, Text Overlay, Custom Scripts)

Built from: comfyui-bundle

Getting Help

• ComfyUI Documentation
• fal.ai Models
• Issues

Related Projects

Project	Purpose
comfyui-bundle	Docker image source code
comfyui-mcp	MCP server for Claude Code
comfyui-genai-connectors	Multi-vendor AI connectors
comfyui-video-utils	Video processing nodes

License

MIT License

---

Made with ❤️ by IO-AtelierTech for creative professionals