█▀▀ █   █▀▀ █  █ █▀▀ █▄  █ █▀▀ █   ▀█▀
█▀▀ █   █▀▀ ▀▄▄▀ █▀▀ █ █ █ █   █    █
▀▀▀ ▀▀▀ ▀▀▀  ▀▀  ▀▀▀ ▀  ▀▀ ▀▀▀ ▀▀▀ ▀▀▀

An inoffical ElevenLabs CLI helper — text-to-speech, sound effects and audiobooks from the command line.

Features

• tts — Convert text to speech with any ElevenLabs voice
• sfx — Generate sound effects from a text prompt
• voices — List and search available voices
• audiobook — Stitch narration, sound effects, and silence into a single audio file

Installation

Homebrew (recommended)

brew tap deegital/tap
brew install elevencli

Go

go install github.com/deegital/elevencli@latest

Build from source

git clone https://github.com/deegital/elevencli.git
cd elevencli
make build

Configuration

An ElevenLabs API key is required. Provide it in one of three ways (highest priority first):

Method	Example
Flag	--api-key sk-...
Environment variable	export ELEVENLABS_API_KEY=sk-...
Config file	~/.elevencli.yaml with api_key: sk-...

Usage

Text-to-Speech

elevencli tts "Hello, world!" --voice JBFqnCBsd6RMkjVDRZzb --output hello.mp3

Flags:

Flag	Default	Description
-v, --voice	(required)	Voice ID
-o, --output	output.mp3	Output file path
-f, --format	mp3	Audio format: mp3, pcm, ulaw
-m, --model	eleven_multilingual_v2	Model ID

Sound Effects

elevencli sfx "heavy rain on a tin roof" --duration 5 --output rain.mp3

Flags:

Flag	Default	Description
-o, --output	output.mp3	Output file path
-d, --duration	auto	Duration in seconds (0.5–30)
-f, --format	mp3	Audio format: mp3, pcm, ulaw

List Voices

elevencli voices
elevencli voices --search "aria"

Flags:

Flag	Description
-s, --search	Filter voices by name (case-insensitive)

Audiobook

Generate a complete audiobook from a JSON script that combines narration, sound effects, and silence:

elevencli audiobook examples/story.json --output story.mp3

Flags:

Flag	Default	Description
-o, --output	audiobook.mp3	Output MP3 file path
--keep-blocks	false	Save individual block audio files

Script Format

The script is a JSON file with an array of blocks. Each block has a type — one of tts, sfx, or silence:

{
  "blocks": [
    {
      "type": "tts",
      "voice": "JBFqnCBsd6RMkjVDRZzb",
      "text": "Once upon a time...",
      "stability": 0.5,
      "similarity_boost": 0.75
    },
    {
      "type": "sfx",
      "text": "gentle ocean waves",
      "background": true,
      "duration": 8.0
    },
    {
      "type": "silence",
      "duration": 1.5
    }
  ]
}

Setting "background": true on an SFX block mixes it with the next TTS block instead of playing sequentially.

Print the full JSON Schema for the script format:

elevencli audiobook schema

See examples/story.json for a complete example.

License

MIT