macOS FFMPEG / Whisper AI video file to transcript script

video-to-transcript.sh

#!/bin/bash

# Video to Audio Transcription Script
# Converts video files to audio and transcribes using Whisper
# Place this file in /usr/local/bin/video-to-transcript.sh
# Usage (global): video-to-transcript.sh myvideo.mp4
# Pre-requisites: Install ffmpeg via brew, install whisper into a venv via pip like:
  # python3.13 -m venv whisper-env
  # source whisper-env/bin/activate
  # pip install -U openai-whisper
  # Create a zshrc alias `alias whisper='(source ~/whisper-env/bin/activate && whisper)'` (not required for this script, just handy)

set -e  # Exit on error

# Check if video file is provided
if [ $# -eq 0 ]; then
    echo "Usage: $0 <video-file>"
    echo "Example: $0 myvideo.mp4"
    exit 1
fi

VIDEO_FILE="$1"

# Check if video file exists
if [ ! -f "$VIDEO_FILE" ]; then
    echo "Error: Video file '$VIDEO_FILE' not found"
    exit 1
fi

# Check if ffmpeg is installed
if ! command -v ffmpeg &> /dev/null; then
    echo "Error: ffmpeg is not installed"
    echo "Install with: brew install ffmpeg"
    exit 1
fi

# Check if whisper virtual environment exists
WHISPER_VENV="$HOME/whisper-env"
if [ ! -d "$WHISPER_VENV" ]; then
    echo "Error: whisper virtual environment not found at $WHISPER_VENV"
    echo "Install with: python -m venv ~/whisper-env && source ~/whisper-env/bin/activate && pip install openai-whisper"
    exit 1
fi

# Get the base filename without extension and the directory
BASENAME=$(basename "$VIDEO_FILE")
BASENAME_NO_EXT="${BASENAME%.*}"
VIDEO_DIR=$(dirname "$VIDEO_FILE")

# Create transcripts folder
TRANSCRIPTS_DIR="${VIDEO_DIR}/${BASENAME_NO_EXT}-transcripts"
mkdir -p "$TRANSCRIPTS_DIR"

AUDIO_FILE="${TRANSCRIPTS_DIR}/${BASENAME_NO_EXT}.mp3"

echo "Converting video to audio..."
echo "Input: $VIDEO_FILE"
echo "Output: $AUDIO_FILE"

# Convert video to audio using ffmpeg
ffmpeg -i "$VIDEO_FILE" -vn -acodec libmp3lame -q:a 2 "$AUDIO_FILE" -y

if [ ! -f "$AUDIO_FILE" ]; then
    echo "Error: Audio conversion failed"
    exit 1
fi

echo "Audio conversion complete!"
echo ""
echo "Starting transcription with Whisper..."

# Activate virtual environment and run whisper with output directory
source "$WHISPER_VENV/bin/activate"
whisper "$AUDIO_FILE" --model turbo --output_dir "$TRANSCRIPTS_DIR"
deactivate

echo ""
echo "Transcription complete!"
echo "All files saved in: $TRANSCRIPTS_DIR"