A Python script that transcribes audio from a video file using OpenAI's Whisper model and saves the transcription as a text file.

transcribe.py

import whisper
import ffmpeg
import os

def main(video_file, language="en"):
    # Load the Whisper model
    print("Loading Whisper model...")
    model = whisper.load_model("base")

    # Paths for the audio and output text filese
    audio_file = "extracted_audio.wav"
    # output_file = os.path.splitext(video_file)[0] + ".txt"
    output_file = video_file + ".txt"

    # Extract audio from the video file
    print("Extracting audio from video file...")
    ffmpeg.input(video_file).output(audio_file).run(overwrite_output=True, quiet=True)

    # Transcribe the extracted audio file
    print("Transcribing audio...")
    result = model.transcribe(audio_file, language=language)

    # Save the transcription to a text file with UTF-8 encoding
    print(f"Saving transcription to {output_file}...")
    with open(output_file, "w", encoding="utf-8") as f:
        f.write(result["text"])

    # Optionally, remove the extracted audio file if not needed
    os.remove(audio_file)
    print("Transcription completed successfully.")

if __name__ == "__main__":
    video_file = "video_to_trancribe.mp4"  # Set video file here
    language = "en"  # Set the desired language code here
    main(video_file, language)