awaz

Text to speech CLI with ElevenLabs voices.

npx awaz

# or

npm i -g awaz

# Usage
awaz "Ship it"

Setup

export ELEVENLABS_API_KEY="your-key"

Get one at elevenlabs.io

Usage

# Speak
awaz "Hello world"

# Pick a voice
awaz -v Roger "Hello world"

# Save to file
awaz -o hello.mp3 "Hello world"

# Pipe it
echo "Hello world" | awaz

# Go faster
awaz --speed 1.2 "Ship faster"

# List voices
awaz voices

Why awaz?

Mac's say command but with voices that don't sound like robots from 2003.

Zero config. One command. Works.

Options

-v, --voice <name>     Voice name or ID
-o, --output <file>    Save to file
-r, --rate <wpm>       Words per minute
--speed <n>            Speed multiplier (0.5-2.0)
--model-id <id>        ElevenLabs model
--stability <n>        Voice consistency (0-1)
--similarity <n>       Match original voice (0-1)
--style <n>            Expressiveness (0-1)

Commands

awaz [text]           # Speak (default)
awaz voices           # List voices
awaz prompting        # Tips for better output

Environment

ELEVENLABS_API_KEY    # Required
ELEVENLABS_VOICE_ID   # Default voice (optional)

For AI Agents

npx skills add ahmadawais/awaz

Agents can use awaz to speak text, generate audio files, or preview how text sounds. See skills/SKILL.md for full agent instructions.

Dev

pnpm install
pnpm build
pnpm test

Inspiration

• macOS say command, the OG.
• steipete's sag in Go.

License

MIT © Ahmad Awais