ElevenLabs Plugin

This plugin provides integration with ElevenLabs text-to-speech services through the ElizaOS platform.

Usage

Add the plugin to your character configuration:

"plugins": ["@elizaos-plugins/plugin-elevenlabs"]

Configuration

The plugin requires these environment variables (can be set in .env file or character settings):

"settings": {
  "ELEVENLABS_API_KEY": "your_elevenlabs_api_key",
  "ELEVENLABS_VOICE_ID": "EXAVITQu4vr4xnSDxMaL",
  "ELEVENLABS_MODEL_ID": "eleven_monolingual_v1",
  "ELEVENLABS_VOICE_STABILITY": "0.5",
  "ELEVENLABS_OPTIMIZE_STREAMING_LATENCY": "0",
  "ELEVENLABS_OUTPUT_FORMAT": "pcm_16000",
  "ELEVENLABS_VOICE_SIMILARITY_BOOST": "0.75",
  "ELEVENLABS_VOICE_STYLE": "0",
  "ELEVENLABS_VOICE_USE_SPEAKER_BOOST": "true"
}

Or in .env file:

ELEVENLABS_API_KEY=your_elevenlabs_api_key
# Optional overrides:
ELEVENLABS_VOICE_ID=EXAVITQu4vr4xnSDxMaL
ELEVENLABS_MODEL_ID=eleven_monolingual_v1
ELEVENLABS_VOICE_STABILITY=0.5
ELEVENLABS_OPTIMIZE_STREAMING_LATENCY=0
ELEVENLABS_OUTPUT_FORMAT=pcm_16000
ELEVENLABS_VOICE_SIMILARITY_BOOST=0.75
ELEVENLABS_VOICE_STYLE=0
ELEVENLABS_VOICE_USE_SPEAKER_BOOST=true

Configuration Options

• ELEVENLABS_API_KEY (required): Your ElevenLabs API credentials.
• ELEVENLABS_VOICE_ID: Optional. Voice selection ID. Defaults to EXAVITQu4vr4xnSDxMaL.
• ELEVENLABS_MODEL_ID: Optional. Speech model ID. Defaults to eleven_monolingual_v1.
• ELEVENLABS_VOICE_STABILITY: Optional. Controls voice stability. Defaults to 0.5.
• ELEVENLABS_OPTIMIZE_STREAMING_LATENCY: Optional. Adjusts streaming latency. Defaults to 0.
• ELEVENLABS_OUTPUT_FORMAT: Optional. Output format (e.g., pcm_16000). Defaults to pcm_16000.
• ELEVENLABS_VOICE_SIMILARITY_BOOST: Optional. Adjusts similarity to the reference voice (0-1). Defaults to 0.75.
• ELEVENLABS_VOICE_STYLE: Optional. Controls voice style intensity (0-1). Defaults to 0.
• ELEVENLABS_VOICE_USE_SPEAKER_BOOST: Optional. Enhances speaker presence (true/false). Defaults to true.
• ELEVENLABS_STT_MODEL_ID: Optional. Speech-to-text model ID. Defaults to scribe_v1.
• ELEVENLABS_STT_LANGUAGE_CODE: Optional. Language code for transcription (e.g., 'en', 'es'). Leave empty for automatic detection.
• ELEVENLABS_STT_TIMESTAMPS_GRANULARITY: Optional. Timestamp detail level: 'none', 'word', or 'character'. Defaults to word.
• ELEVENLABS_STT_DIARIZE: Optional. Enable speaker diarization (true/false). Defaults to false.
• ELEVENLABS_STT_NUM_SPEAKERS: Optional. Expected number of speakers for diarization (1-32).
• ELEVENLABS_STT_TAG_AUDIO_EVENTS: Optional. Tag audio events like laughter, applause (true/false). Defaults to false.

The plugin provides the following model types:

• TEXT_TO_SPEECH: Converts text into spoken audio.
• TRANSCRIPTION: Converts audio/video into text transcripts.