Google Cloud TTS

Google Cloud Text-to-Speech is a text-to-speech plugin. It synthesizes the LLM’s reply into the agent’s voice using Google Cloud’s voice families, from Standard through Chirp 3 HD. It occupies the pipeline’s tts slot.

Setup

Set your Google credentials in the worker environment. Create a service account in the Google Cloud console and point the SDK at its JSON key. The Python SDK resolves credentials in priority order: credentials_json → service_account_path → api_key → the GOOGLE_APPLICATION_CREDENTIALS env var → a credentials.json file in the working directory.

export GOOGLE_APPLICATION_CREDENTIALS=/path/to/service-account.json

The api_key argument also falls back to the GOOGLE_API_KEY env var.

Usage

Import the plugin and pass it to the pipeline’s tts slot.

from zrt.plugins import GoogleTTS

tts = GoogleTTS(
    voice="en-US-Neural2-F",
    language_code="en-US",
)

# Pipeline(tts=tts, ...)

Voice families

Cloud TTS routes by voice name, so the voice family is normally inferred from voice. You can optionally pin the family with model, which accepts one of the known GoogleTTSModel values: "chirp3-hd", "chirp-hd", "studio", "neural2", "wavenet", "standard", "news", "polyglot", "chirp". If the explicit model contradicts the family implied by voice, a ValueError is raised.

Python

from zrt.plugins import GoogleTTS

tts = GoogleTTS(
    voice="en-US-Chirp3-HD-Charon",
    model="chirp3-hd",
)

Parameters

Constructor parameters for the Python SDK (GoogleTTS). The JavaScript and Go SDKs expose voice, languageCode, model, speakingRate/SpeakingRate, and pitch in their idiomatic form.

Parameter	Type	Default	Description
`api_key`	`str`	`None`	Google API key. Also used as a fallback credential source; falls back to the `GOOGLE_API_KEY` env var.
`credentials_json`	`str`	`None`	Service-account credentials as a raw JSON string or a path to a JSON key file. Highest precedence.
`service_account_path`	`str`	`None`	Path to a service-account JSON key file. Used when `credentials_json` is not provided.
`stream`	`bool`	`True`	Stream audio back progressively for lower time-to-first-audio.
`voice`	`str`	`"en-US-Neural2-F"`	Voice name in the form `{language}-{region}-{Family}-{Character}`. The name implicitly determines the voice family.
`language_code`	`str`	`"en-US"`	BCP-47 language code for the voice.
`model`	`str`	`""`	Voice family override (`GoogleTTSModel`). Leave empty to let the voice name determine the family.
`sample_rate`	`int`	`24000`	Output audio sample rate in Hz.
`speaking_rate`	`float`	`None`	Playback speed multiplier, range `[0.25, 4.0]`. `None` uses the API default.
`pitch`	`float`	`None`	Pitch shift in semitones, range `[-20.0, 20.0]`. `None` omits the parameter.
`voice_config`	`GoogleVoiceConfig`	`None`	Optional voice-selection dataclass (`name`, `languageCode`). Its fields take precedence over `voice` / `language_code` when those are still at their defaults.

Import paths

SDK	Import	Constructor
Python	`from zrt.plugins import GoogleTTS`	`GoogleTTS(...)`
JavaScript	`from '@zrt/js-sdk/plugins/google'`	`GoogleTTS({ ... })`
Go	`.../zrt-golang-sdk/plugins/google`	`google.NewTTS(google.TTSOptions{...})`

The synthesized audio is streamed back to the caller as the final stage of the pipeline.

Plugins

Inference Gateway

Speech-to-text

LLM

Text-to-speech

Realtime models

Avatars

Turn detection

Voice activity detection

Noise cancellation

Setup

Usage

Voice families

Parameters

Import paths

​Setup

​Usage

​Voice families

​Parameters

​Import paths

Setup

Usage

Voice families

Parameters

Import paths