How to beginner · 3 min read

How to build meeting transcription app with Whisper

Q: How to build meeting transcription app with Whisper

Use OpenAI's whisper-1 model via the openai Python SDK to transcribe meeting audio files. Upload audio in supported formats (mp3, wav, m4a), call client.audio.transcriptions.create(), and extract the text transcript for your app.

Quick answer

Use OpenAI's whisper-1 model via the openai Python SDK to transcribe meeting audio files. Upload audio in supported formats (mp3, wav, m4a), call client.audio.transcriptions.create(), and extract the text transcript for your app.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0

Setup

Install the openai Python package and set your OpenAI API key as an environment variable.

Install package: pip install openai
Set environment variable in your shell:
export OPENAI_API_KEY='your_api_key_here'

bash

pip install openai

Step by step

This example shows how to transcribe a local meeting audio file using the Whisper API. It uploads the file, calls the transcription endpoint, and prints the transcript text.

python

import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

def transcribe_meeting(audio_path: str) -> str:
    with open(audio_path, "rb") as audio_file:
        transcript = client.audio.transcriptions.create(
            model="whisper-1",
            file=audio_file
        )
    return transcript.text

if __name__ == "__main__":
    audio_file_path = "meeting_recording.mp3"  # Replace with your audio file path
    text = transcribe_meeting(audio_file_path)
    print("Transcription output:")
    print(text)

output

Transcription output:
Good morning everyone, let's start the meeting with the project updates...

Common variations

You can use different audio formats supported by Whisper such as WAV, M4A, or MP4. For real-time transcription, stream audio chunks and call the API incrementally (requires custom buffering). You can also use the openai-whisper local package for offline transcription.

python

import asyncio
from openai import OpenAI

client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

async def async_transcribe(audio_path: str):
    with open(audio_path, "rb") as f:
        transcript = await client.audio.transcriptions.acreate(
            model="whisper-1",
            file=f
        )
    print(transcript.text)

if __name__ == "__main__":
    asyncio.run(async_transcribe("meeting_recording.wav"))

output

Good morning everyone, let's start the meeting with the project updates...

Troubleshooting

If you get a FileNotFoundError, verify the audio file path is correct.
If transcription is inaccurate, ensure audio quality is clear and in a supported format.
For AuthenticationError, check your OPENAI_API_KEY environment variable is set properly.
Large audio files may require chunking or trimming to fit API limits.

✅

Key Takeaways

Use OpenAI's whisper-1 model with the official openai Python SDK for meeting transcription.
Supported audio formats include mp3, wav, m4a, and mp4; ensure good audio quality for best results.
Set your API key securely via environment variables and handle large files by chunking if needed.

Verified 2026-04 · whisper-1

Verify ↗