A high-performance, multi-threaded asset downloader for ComfyUI. This suite is designed to mimic the Linux wget experience on Windows, providing simultaneous model downloading with real-time progress tracking and auto-resume capabilities.
Before downloading, the system needs a place to store the massive 20GB of data.
- Purpose: Creates the local directory structure for ComfyUI.
- Action: Generates
models/checkpoints,models/clip, andmodels/vaefolders. - Why: Prevents write errors by ensuring the destination exists before the download starts.
Once the folders are ready, this is your primary entry point for asset acquisition.
- Purpose: Acts as a bridge between Windows and the Python Download Engine.
- Features:
- Auto-Dependency: Checks for Python and automatically installs the
requestslibrary. - Simultaneous Streams: Downloads the Checkpoint, CLIP, and VAE at the same time.
- Wget Aesthetics: Displays authentic
[=====> ]progress bars. - Fault Tolerance: If your internet blips, simply run this file again to resume.
- Auto-Dependency: Checks for Python and automatically installs the
| File Name | Category | Size | Destination |
|---|---|---|---|
| z_image_turbo_bf16 | Checkpoint | 11.4 GB | /models/checkpoints/ |
| qwen_3_4b | Text Encoder | 7.0 GB | /models/clip/ |
| z_image_ae | VAE | 0.2 GB | /models/vae/ |
- Place all repository files in your desired installation directory.
- Run
install_comfyui.batonce. - Run
z-image.batto begin the download. - Once the progress bars hit 100%, your ComfyUI instance is ready for Z-Image Turbo.
- Python 3.10+ (Must be added to System PATH)
- Storage: ~20GB of free disk space.
- Internet: A stable connection is recommended for large binary files.
Maintained by AB HASH.