lyria-cli

Generate music with Google Lyria from your terminal or an AI agent. Text in, MP3 out, with JSON results for automation.

Website · GitHub

Quick start

Requires Node.js 22+ and a Gemini API key. After the first npm release:

npm install -g @crapthings/lyria-cli
export GEMINI_API_KEY='your-api-key'

lyria-cli generate \
  --prompt 'Gentle jazz piano and upright bass, no vocals' \
  --output ./music.mp3 \
  --json

The npm package is @crapthings/lyria-cli; the executable is lyria-cli. Use --prompt-file ./song.txt for lyrics or long prompts. Run lyria-cli --help for all options. Default model: lyria-3.5.

Agent skill

The npm package includes lyria-music. After a project-local npm install @crapthings/lyria-cli, register it with your agent:

npx skills add ./node_modules/@crapthings/lyria-cli/skills/lyria-music

From this checkout, use npx skills add ./skills/lyria-music. You can also install from GitHub with npx skills add crapthings/lyria-cli --skill lyria-music. Installing the skill does not install the CLI or API credentials.

Development

pnpm install
pnpm check
pnpm test
node dist/cli.js --help

Release a new version with npm run release (patch by default). Preview with npm run release -- --dry-run.

CLI reference · Development, releases & GitHub Pages

MIT