This repository provides an extensive suite of over 210 custom nodes for ComfyUI, aimed at enhancing image, text, and video processing capabilities. It offers advanced functionalities for users looking to expand ComfyUI's workflow potential, including complex image manipulations, dynamic text prompting, and experimental video generation.

How It Works

The suite integrates a wide array of specialized nodes, leveraging libraries like matplotlib and scikit-learn for image analysis and manipulation. Key features include advanced masking operations, image blending modes, noise generation, and sophisticated text processing for dynamic prompting and file handling. It also introduces experimental video nodes and integrates with models like BLIP and MiDaS for enhanced image understanding and depth approximation.

Quick Start & Requirements

• Installation: Clone the repository into your ComfyUI custom nodes directory. Install dependencies using pip install -r requirements.txt within the cloned directory, preferably using ComfyUI's embedded Python.
• Prerequisites: Python, ComfyUI. Video nodes require FFmpeg and potentially OpenH264. Some image nodes have optional dependencies like matplotlib, scikit-learn, and pilgram. The Image Crop Face node requires a CUDA-enabled GPU.
• Setup Time: Estimated 5-15 minutes for cloning and dependency installation.
• Documentation: Salt AI Documentation

Highlighted Details

• Over 210 specialized nodes for image, text, and video manipulation.
• Advanced text processing with support for A1111 wildcards and NSP prompts.
• Experimental video nodes (Write to Video, Create Video from Path) requiring FFmpeg.
• Image processing nodes include advanced masking, blending, filters, and face cropping (GPU required).
• Caching nodes for latent tensors and conditioning to disk.

Maintenance & Community

• Development Status: The project is not under active development as of December 15, 2023. The author is open to forks and PRs.
• Community: No specific community links (Discord/Slack) are provided in the README.

Licensing & Compatibility

• License: The README does not explicitly state a license. It is hosted on GitHub, implying a default open-source license, but specific terms are not detailed.
• Compatibility: Designed for ComfyUI. Video nodes have specific Windows installation notes.

Limitations & Caveats

• The project is not under active development, meaning future updates or bug fixes are not guaranteed.
• Video nodes are experimental and have specific installation requirements and known issues (e.g., FFV1 container warnings).
• The Image Crop Face node has limitations regarding face detection and requires a GPU.
• Some nodes require manual installation of external libraries or binaries.