@skxv/transcribe

A terminal UI for transcribing local audio files with AssemblyAI.

Install and run

npx @skxv/transcribe

Requires Bun (recommended) or Node.js 26.1+. On older Node versions, the CLI automatically re-runs via Bun when it is on your PATH.

For local development:

bun install
bun run dev

First run

On first launch, the app prompts for your AssemblyAI API key in the terminal. The key is saved to your user config directory (~/.config/@skxv/transcribe/config.json on macOS/Linux). The app does not read .env files or environment variables for the API key.

Use

Start the TUI, drag one or more audio files into the terminal input, then press Enter. Most terminals paste dragged files as paths, which the app validates and sends to AssemblyAI.

Completed transcripts are saved to ~/Documents/Transcriptions/<audio-file-name>.txt. Select a past transcript in the History panel (or press H to focus it, then Enter) to open it again.

Keyboard shortcuts

| Key | Action | | ----------------------- | ---------------------------------------------------- | | Tab | Next settings control | | H | Focus history | | Enter (history focused) | Open selected transcript | | Ctrl+O | Reveal transcript or audio file in Finder / Explorer | | N / P | Next / previous utterance | | E | Edit utterance line | | R | Rename speaker | | Space / X | Play/pause / stop audio | | [ / ] | Seek backward / forward 10 seconds |

Transcription mode

Dropped files are transcribed with AssemblyAI's pre-recorded audio API using universal-3-pro with universal-2 fallback.

Publish

npm run build
npm publish --access public