Transcribe a hosted audio file with Deepgram Nova-3 in Python

What you get

A Python script that transcribes a remote audio file with the nova-3 model and smart formatting.

Prerequisites

Python 3
A Deepgram API key exported as DEEPGRAM_API_KEY (or in a .env file)

Steps

Install the SDK:

Shell

pip install deepgram-sdk python-dotenv

Create transcribe.py:

Python

import os
import dotenv
from deepgram import DeepgramClient

dotenv.load_dotenv()

def main():
    try:
        deepgram = DeepgramClient(api_key=os.getenv("DEEPGRAM_API_KEY"))
        response = deepgram.listen.v1.media.transcribe_url(
            url="https://dpgr.am/bueller.wav",
            model="nova-3",
            smart_format=True,
        )
        print(f"response: {response}\n\n")
    except Exception as e:
        print(f"Exception: {e}")

if __name__ == "__main__":
    main()

Test it

Run python transcribe.py. The printed response contains the transcript of the sample Ferris Bueller clip; replace url with your own publicly reachable audio file.