Skip to content

First Run

After installation, run corridorkey init once to complete setup. This command checks your environment, creates the config file, and downloads the inference model.

corridorkey init

What Init Does

Init runs three steps in order:

  1. Environment check - runs corridorkey doctor and prints a results table showing Python version, FFmpeg, GPU, and model status.
  2. Config file - creates ~/.config/corridorkey/corridorkey.yaml with default settings if it does not already exist.
  3. Model download - checks whether the inference model is present and offers to download it if not.

The Environment Check Table

The table shows a row for each check with a status of OK, WARN, or FAIL.

FAIL on any of these three means CorridorKey will not work until fixed:

  • Python >= 3.13 - the installed Python version is too old
  • ffmpeg - FFmpeg is not installed or not on PATH
  • inference model - the model file has not been downloaded yet

WARN rows (git, config file, checkpoint directory) do not block processing. Init will continue past them.

Downloading the Model

When init reaches the model step it shows the download URL and asks for confirmation:

Inference model not found in: ~/.config/corridorkey/models
URL: https://huggingface.co/...
Download inference model now? [y/n] (y):

The model is approximately 400 MB. A progress bar shows download progress. The file is saved to ~/.config/corridorkey/models/ and verified against a checksum.

If you decline the download, init prints the manual instructions:

To download manually, place greenformer_v2.pth in:
  ~/.config/corridorkey/models
Then run corridorkey doctor to verify.

After Init

When init completes successfully it prints:

Init complete. Run `corridorkey wizard` to get started.

Run corridorkey doctor at any time to re-check the environment without modifying anything.