Google STT

The GoogleSTTService provides real-time speech recognition using Google Cloud's Speech-to-Text V2 API. It supports streaming audio, multi-language detection, and automatic reconnection to handle Google's 5-minute streaming limit.

Installation

To use Google STT, install the required dependencies:

pip install "piopiy-ai[google]"

Prerequisites

• A Google Cloud Project with the Speech-to-Text API enabled.
• Google Cloud credentials (JSON key file) (Get yours here).
• Set the credentials environment variable:

  export GOOGLE_APPLICATION_CREDENTIALS="path/to/your/service-account-file.json"

Configuration

GoogleSTTService Parameters

Parameter	Type	Default	Description
credentials	str	None	JSON string of service account credentials.
credentials_path	str	None	Path to service account JSON file.
location	str	"global"	Google Cloud location (e.g., "us-central1").
sample_rate	int	None	Audio sample rate in Hz.
params	InputParams	InputParams()	Advanced recognition settings.

InputParams

Parameter	Type	Default	Description
languages	List[Language]	[EN_US]	Recognition languages. First is primary.
model	str	"latest_long"	Speech recognition model to use.
enable_automatic_punctuation	bool	True	Add punctuation to transcripts.
enable_interim_results	bool	True	Stream partial recognition results.
enable_voice_activity_events	bool	False	Detect voice activity in audio.

Usage

Basic Setup

from piopiy.services.google.stt import GoogleSTTService
from piopiy.transcriptions.language import Language

stt = GoogleSTTService(
    credentials_path="path/to/creds.json",
    params=GoogleSTTService.InputParams(
        languages=[Language.EN_US, Language.ES_ES],
        model="latest_long"
    )
)

Notes

• Streaming Limit: Google STT has a 5-minute limit per stream. Piopiy automatically handles reconnection to enable "endless streaming."
• Model Choice: latest_long is recommended for general conversational use cases.