Google TTS

The Google TTS plugin enables your agent to use Google's text-to-speech models for generating natural-sounding voice output. It supports low-latency gRPC streaming with Chirp 3 HD voices and Vertex AI endpoints.

Installation

pip install "videosdk-plugins-google"

Authentication

Set your Google API key as an environment variable:

export GOOGLE_API_KEY="your-google-api-key"

You can obtain an API key from the Google AI Studio.

Example Usage

from videosdk.plugins.google import GoogleTTS, GoogleVoiceConfig
from videosdk.agents import CascadingPipeline

# Configure voice settings
voice_config = GoogleVoiceConfig(
    languageCode="en-US",
    name="en-US-Chirp3-HD-Aoede",
    ssmlGender="FEMALE"
)

# Initialize the Google TTS model
tts = GoogleTTS(
    # When GOOGLE_API_KEY is set in .env - DON'T pass api_key parameter
    api_key="your-google-api-key",
    speed=1.0,
    pitch=0.0,
    voice_config=voice_config,
    custom_pronunciations=[{"tomato": "təˈmeɪtoʊ"}],  # Optional IPA overrides
)

# Add tts to cascading pipeline
pipeline = CascadingPipeline(tts=tts)

Vertex AI

To use the Vertex AI endpoint instead of an API key, authenticate using Application Default Credentials (ADC) and set your project ID:

export GOOGLE_CLOUD_PROJECT="my-gcp-project"

from videosdk.plugins.google import GoogleTTS, VertexAIConfig

tts = GoogleTTS(
    vertexai=True,
    vertexai_config=VertexAIConfig(location="us-central1"),
    streaming=False,  # Streaming cannot be used with Vertex AI
)

note

• streaming=True (the default) requires a Chirp 3 HD voice (e.g. en-US-Chirp3-HD-Aoede) and cannot be combined with vertexai=True.
• Vertex AI requires a GCP project ID via VertexAIConfig(project_id="..."), the GOOGLE_CLOUD_PROJECT env variable, or a GOOGLE_APPLICATION_CREDENTIALS service-account file.

Configuration Options

• api_key: (str) Your Google Cloud TTS API key. Can also be set via the GOOGLE_API_KEY environment variable.
• speed: (float) The speaking rate of the generated audio (default: 1.0).
• pitch: (float) The pitch of the generated audio. Can be between -20.0 and 20.0 (default: 0.0).
• response_format: (str) The format of the audio response. Currently only supports "pcm" (default: "pcm").
• voice_config: (GoogleVoiceConfig) Configuration for the voice to be used.
  • languageCode: (str) The language code of the voice (e.g., "en-US", "en-GB") (default: "en-US").
  • name: (str) The name of the voice to use (e.g., "en-US-Chirp3-HD-Aoede", "en-US-News-N") (default: "en-US-Chirp3-HD-Aoede").
  • ssmlGender: (str) The gender of the voice ("MALE", "FEMALE", "NEUTRAL") (default: "FEMALE").
• custom_pronunciations: (list[dict] | dict | None) IPA pronunciation overrides for specific words (e.g., [{"tomato": "təˈmeɪtoʊ"}]). Defaults to None.
• streaming: (bool) Use gRPC StreamingSynthesize for lower-latency audio generation. Only compatible with Chirp 3 HD voices and cannot be combined with vertexai=True (default: True).
• vertexai: (bool) Use the Vertex AI TTS endpoint with Application Default Credentials (ADC) instead of an API key (default: False).
• vertexai_config: (VertexAIConfig) Project and region settings for Vertex AI.
  • project_id: (str | None) Your GCP project ID. Falls back to GOOGLE_CLOUD_PROJECT or GOOGLE_APPLICATION_CREDENTIALS (default: None).
  • location: (str) GCP region for the TTS endpoint (default: "us-central1").

Additional Resources

The following resources provide more information about using Google with VideoSDK Agents SDK.

• Google TTS docs: Google Cloud TTS documentation.
• Chirp 3 HD voices: Available voices for low-latency streaming synthesis.
• Vertex AI TTS: Vertex AI Text-to-Speech documentation.

SDK Reference

GitHub Repository

Python Package