DAOTUI

A terminal UI for transcribing audio on an Azure GPU VM with WhisperX: it starts the VM, uploads your file, runs transcription, pulls the result back, lets you QC low-confidence words, then cleans up and deallocates the VM.

Setup

1. External tools

These must be installed and on your PATH — the app shells out to them, so npm install does not provide them:

Tool	Purpose	Install
Node.js ≥ 18	Runs the app	nodejs.org or brew install node
Azure CLI (az)	Starts / deallocates the VM	brew install azure-cli
ssh, scp, ssh-keygen	Upload, run, download, cleanup, key generation	Preinstalled on macOS (OpenSSH)

Then authenticate the Azure CLI once:

az login

2. Run the app

From the repo root:

npm install
npm start

3. Verify your environment

node --version    # >= 18
az --version      # Azure CLI present
az account show   # logged in
ssh -V            # present (preinstalled on macOS)

Configuration

Settings and the SSH key path are stored outside the repo, in the OS config directory (~/Library/Preferences/daotui-nodejs/config.json).