ElevenLabs (Premium TTS)

ElevenLabs provides state-of-the-art neural text-to-speech with natural-sounding voices, voice cloning, and multi-language support.

When to use ElevenLabs

• You want the most natural-sounding voice output
• You need voice cloning (replicate a specific voice)
• You want multi-lingual TTS (Cantonese, Mandarin, English, etc.)
• The free Edge TTS voice quality is insufficient

Installation

The ElevenLabs extra is included in the [all] install. If you installed the minimal bundle:

cd ~/.monoclaw/monoclaw-runtime
uv pip install -e ".[tts-premium]"

Setup

1. Sign up at elevenlabs.io
2. Copy your API key from the dashboard
3. Configure MonoClaw:

monoclaw config set ELEVENLABS_API_KEY "your-key"
monoclaw config set tts.provider elevenlabs

Or in config.yaml:

tts:
  provider: elevenlabs
  elevenlabs:
    voice_id: "21m00Tcm4TlvDq8ikWAM"  # Default: Rachel
    model: "eleven_multilingual_v2"

Voice selection

Browse available voices at elevenlabs.io/voice-library. Popular choices:

Set a voice:

monoclaw config set tts.elevenlabs.voice_id "AZnzlk1XvdvUeBnXmlld"

Voice cloning

You can clone your own voice:

1. Go to elevenlabs.io/voice-lab
2. Upload 1–30 minutes of clean audio
3. Copy the new voice ID into your config

Using TTS

In the CLI or any messaging platform:

/voice on

Mona will speak her replies. Press Ctrl+B in the CLI to record a voice message.

Cost

ElevenLabs charges per character:

• ~$0.003 per 1K characters on the Starter plan
• Free tier: 10K characters/month

Monitor usage in the ElevenLabs dashboard.

Comparison with Edge TTS

Troubleshooting