Installation
CorridorKey is installed as a command-line tool using the one-line installers below. The installer handles Python, the package manager, and the tool itself.
System Requirements
| Requirement | Minimum |
|---|---|
| Operating system | Windows 10, macOS 12, or Ubuntu 20.04 |
| Python | 3.13 (installed automatically if missing) |
| Disk space | 2 GB (500 MB tool + ~1.4 GB model) |
| RAM | 8 GB |
| GPU | Optional but strongly recommended (see below) |
GPU recommendations
Without a GPU, inference runs on CPU and is very slow (several minutes per frame). A GPU is strongly recommended for any practical use.
| Platform | Recommended |
|---|---|
| Windows / Linux | NVIDIA GPU with 4 GB VRAM or more (CUDA) |
| macOS Apple Silicon | M1 or later (MLX) |
| macOS Intel | CPU only |
Windows
Open PowerShell and run:
irm https://corridorkey.dev/install.ps1 | iex
The installer will ask which GPU you have, install uv if needed, install CorridorKey, run first-time setup, and create a CorridorKey - Drop Clips Here.bat shortcut on your Desktop.
macOS and Linux
Open Terminal and run:
curl -sSf https://corridorkey.dev/install.sh | bash
The installer detects Apple Silicon automatically and selects the MLX build. On Linux it asks whether you have an NVIDIA GPU. After setup it creates a launcher on your Desktop.
Manual Installation
If you prefer to install without the script, use uv tool install directly.
For NVIDIA GPU (Windows/Linux):
uv tool install "corridorkey-cli[cuda]" --python 3.13
For Apple Silicon (macOS):
uv tool install "corridorkey-cli[mlx]" --python 3.13
CPU only:
uv tool install corridorkey-cli --python 3.13
uv must be installed first. See the uv installation guide.
After manual installation, run corridorkey init to complete setup.
Verifying the Installation
corridorkey --help
If the command is not found after installation, close and reopen your terminal to pick up the updated PATH.