Transcribe audio files with Whisper Large v3 Turbo on Groq

What you get

Fast audio-to-text transcription using Groq-hosted whisper-large-v3-turbo from a Python script.

Prerequisites

A Groq API key in GROQ_API_KEY
Python with the groq package
An audio file (FLAC, MP3, MP4, MPEG, MPGA, M4A, OGG, WAV or WebM; up to 25 MB on the free tier)

Steps

Install the SDK and set your key.

Shell

pip install groq
export GROQ_API_KEY="your-api-key"

Create transcribe.py next to your audio file.

Python

import os
from groq import Groq

client = Groq()
filename = os.path.dirname(__file__) + "/YOUR_AUDIO.wav"

with open(filename, "rb") as file:
    transcription = client.audio.transcriptions.create(
      file=file,
      model="whisper-large-v3-turbo",
      prompt="Specify context or spelling",
      response_format="verbose_json",
      language="en",
      temperature=0.0
    )
    print(transcription.text)

Test it

Shell

python transcribe.py

The transcript prints to your terminal. Use whisper-large-v3 instead if you need higher accuracy over speed.