Share AI Models Between Programs with Symbolic Links

How to use the same model without duplicating it between different programs? (Forge, ComfyUI, InvokeAI, Fooocus, SwarmUI, Kohya_SS...) (Windows)

Short answer: Use the symbolic link feature. (Creates a 0 KB duplicate file that points to the original)

Example Structure

This is how you can store one large model file and have multiple programs use it without taking up extra disk space:


- D:\AI_MODELS\ (Your central storage)
  └── sd_xl_base_1.0.safetensors (12.8 GB)

- C:\ComfyUI\models\checkpoints\
  └── sd_xl_base_1.0.safetensors (Symbolic Link, 0 KB)

- C:\Fooocus\models\checkpoints\
  └── sd_xl_base_1.0.safetensors (Symbolic Link, 0 KB)

1. Installation

  1. Download a Link Shell Extension.
  2. Install - HardLinkShellExt_X64.exe.

2. Usage

  1. Find the model file you want to share.
  2. Right-click on the model > "Show more options" (on Windows 11) > Pick Link Source.
  3. Go to the folder where you want to place the "shortcut" (e.g., inside another program's models folder).
  4. Right-click in an empty space > "Show more options" (on Windows 11) > Drop As... > Symbolic link.
  5. Ready! A link file will be created, which the program will read as if it were the original.

3. Important Note!

Do not copy the link file to another location. If you try to copy and paste the symbolic link file itself, Windows will copy the *original* large file, defeating the purpose.

Explanation

Symbolic Links (Symlinks)

Think of a symbolic link as a smart shortcut. It's a tiny file that contains only the path to the original file or folder. The key advantage is that it can point to items on different drives (e.g., a model on `D:\` can be linked into a program's folder on `C:\`). This is the most flexible and recommended method for sharing models between different AI programs.

Hard Links

A hard link is a direct reference to the data on the disk. It's like having two file names pointing to the exact same data. The data is only deleted when the very last hard link is removed. However, hard links have a major limitation: they cannot point to folders and must be on the same drive/partition as the original file. While robust, they are less versatile than symbolic links for this use case.