ComfyUI-KJNodes 🧩 Workflow Open source

What it is

ComfyUI-KJNodes is a large collection of utility nodes by Kijai (GPL-3.0) that many popular video and image workflows depend on. It adds nodes for masks, image batches, resizing, text, conditioning, model patches, VRAM management, and helpers used in Wan, LTX and other video pipelines. If you load a shared workflow and see missing "KJ" nodes, this is the pack. Kijai also maintains the WanVideo and other wrapper repos, and KJNodes often receives helper nodes the same week a new open model ships, which is why so many community video templates list it as a requirement.

Key features

• Mask creation, growing and batch utilities
• Image resize, crop, concatenate and batch tools
• Model patches and speed/memory helpers used in video workflows
• Set/Get nodes to tidy large graphs
• Text and string utilities for prompt handling
• Frequently updated alongside new model releases

How to install

Clone into ComfyUI/custom_nodes and run pip install -r requirements.txt, or install "KJNodes" through ComfyUI-Manager.

Pricing and rights

Free, GPL-3.0. Nodes add no watermark or usage limit; output rights and commercial use follow the licences of the models you load, not the node pack.

Who it is for

Anyone running community video or image workflows in ComfyUI, and workflow builders who want cleaner graphs with fewer wires. It is also useful for batch jobs, such as resizing and masking many product images or video frames in one run.

Verdict

An unglamorous but near-mandatory dependency pack, maintained actively. Documentation is thin, so you often learn nodes from example workflows.