Skip to content

Text2Video workflow using ComfyUI with Wan2.1 14B FusionX LoRA, UMT5 encoder, and FP8 VAE. Includes custom nodes, Docker-ready setup, and scripts to download models for reproducible video generation from text prompts.

License

Notifications You must be signed in to change notification settings

pfrancug/text2video-wan2.1-fusionx

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🚀 text2video-wan2.1-fusionx

Text2Video workflow powered by ComfyUI, Wan2.1 14B FusionX LoRA, UMT5 encoder, and FP8 VAE.
Generate high-quality videos from text prompts with reproducible results, custom nodes, and a Docker-ready setup.


✨ Features

  • ComfyUI Integration: Automated setup and updates via PowerShell scripts.
  • Wan2.1 14B FusionX LoRA: Advanced LoRA for text-to-video generation.
  • UMT5 Encoder & FP8 VAE: Enhanced encoding and video quality.
  • Custom Nodes: Extend ComfyUI with VideoHelperSuite and WanVideoWrapper.
  • Docker Support: GPU/CPU containers for easy deployment.
  • Model Management: Scripts to download and organize models.

🗂️ Project Structure

📁 docker/
│   ├── comfy.docker-compose.yml        # Docker Compose for ComfyUI
│   ├── comfy.dockerfile                # Dockerfile for ComfyUI
│   ├── cuda.docker-compose.yml         # Docker Compose for CUDA
│   └── cuda.dockerfile                 # Dockerfile for CUDA
📁 powershell/
│   ├── cuda-build.ps1                  # Build CUDA base image
│   ├── comfy-run.ps1                   # Start/stop/rebuild ComfyUI Docker
│   ├── prepare-comfy.ps1               # Download & setup ComfyUI
│   ├── prepare-video_helper_suite.ps1  # Setup VideoHelperSuite node
│   └── prepare-wan_video_wrapper.ps1   # Setup WanVideoWrapper node
📁 tools/
│   └── overrides/                      # Directory for custom nodes, models, packages, and user configs

⚡ Quickstart

1. Clone the Repository

git clone https://github.com/pfrancug/text2video-wan2.1-fusionx.git
cd text2video-wan2.1-fusionx

2. Prepare ComfyUI & Custom Nodes

Run the PowerShell scripts (Windows):

powershell\prepare-comfy.ps1
powershell\prepare-video_helper_suite.ps1
powershell\prepare-wan_video_wrapper.ps1

3. Start with Docker

powershell\cuda-build.ps1
powershell\comfy-run.ps1

📦 Dependencies

This workflow relies on the following software and resources:


📝 License

MIT License © 2025 Piotr Francug
See LICENSE for details.


💡 Tips

  • For latest ComfyUI, run powershell\prepare-comfy.ps1 --latest
  • Use powershell\comfy-run.ps1 to rebuild and restart containers safely.
  • All scripts run relative to their own directory for reliability.

🛠️ Roadmap / TODO

  • Implement automated model downloads with curl.
  • Adjust workflow JSON files for reproducibility.
  • Create a main prepare.ps1 script to orchestrate all setup tasks.
  • Transform all PowerShell scripts into cross-platform .sh versions.
  • Validate full setup from environment preparation to video generation.

About

Text2Video workflow using ComfyUI with Wan2.1 14B FusionX LoRA, UMT5 encoder, and FP8 VAE. Includes custom nodes, Docker-ready setup, and scripts to download models for reproducible video generation from text prompts.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published