ElevenLabs

The ElevenLabsTTSService provides the highest quality AI voices with emotional range and personality. It's the recommended TTS provider for premium, high-fidelity voice applications.

Installation

To use ElevenLabs, install the required dependencies:

pip install "piopiy-ai[elevenlabs]"

Prerequisites

• An ElevenLabs account and API key (Get yours here).
• Set your API key in your environment:

  export ELEVENLABS_API_KEY="your_api_key_here"

Configuration

ElevenLabsTTSService Parameters

Parameter	Type	Default	Description
api_key	str	Required	Your ElevenLabs API key.
voice_id	str	"pNInz6obpgmqMHCicJ4A"	Voice ID (e.g., "Rachel").
model	str	"eleven_turbo_v2_5"	Model ID for synthesis.
output_format	str	"pcm_16000"	Audio format (e.g., "pcm_16000").

Usage

Basic Setup

import os
from piopiy.services.elevenlabs.tts import ElevenLabsTTSService

tts = ElevenLabsTTSService(
    api_key=os.getenv("ELEVENLABS_API_KEY"),
    voice_id="pNInz6obpgmqMHCicJ4A", # Rachel
    model="eleven_turbo_v2_5"
)

Notes

• Latency: eleven_turbo_v2_5 is recommended for real-time voice agents due to its optimized latency and high quality.
• Custom Voices: You can use your own cloned voices by providing their unique voice_id.
• WebSocket Streaming: This service uses ElevenLabs' WebSocket API to stream audio as it's generated.