Groq (Whisper)

Overview

The GroqSTTService provides ultra-fast transcription using Groq's high-speed LPU™ Inference Engine running Whisper models. It is ideal for applications where extremely low latency is critical.

Installation

To use Groq STT, install the required dependencies:

pip install "piopiy-ai[groq]"

Prerequisites

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

  export GROQ_API_KEY="your_api_key_here"

Configuration

GroqSTTService Parameters

Parameter	Type	Default	Description
api_key	str	None	Groq API key (defaults to env var).
model	str	"whisper-large-v3-turbo"	Whisper model ID.
language	Language	EN	Language of the audio input.
prompt	str	None	Optional text to guide the model style.
temperature	float	0.0	Sampling temperature for transcription.

Usage

Basic Setup

import os
from piopiy.services.groq.stt import GroqSTTService
from piopiy.transcriptions.language import Language

stt = GroqSTTService(
    api_key=os.getenv("GROQ_API_KEY"),
    model="whisper-large-v3-turbo",
    language=Language.EN
)

Notes

• Performance: Groq's specialized hardware allows for near-instantaneous transcription using large Whisper models.
• Segmented Processing: This service processes audio in segments, typically yielding results at the end of a user's utterance.