download-youtube-transcript

download-youtube-transcript.sh

#!/bin/sh

# We expect a single parameter, the URL of the YouTube video
if [ $# -ne 1 ]; then
  echo "Usage: $0 <YouTube URL>"
  exit 1
fi

# Get a temp file name
tempfile=$(mktemp)

LANG="en"
OUTPUT_FILE="$tempfile.$LANG.vtt"

# Remove the temp file on exit
trap "rm -f $OUTPUT_FILE" EXIT

youtube-dl -o "$tempfile" --write-sub --sub-lang $LANG --skip-download "$1" 2>/dev/null >/dev/null

# Extract the transcript from the subtitle file.
# We remove lines with "-->" and the first 5 lines.
TEXT=$(
    cat "$OUTPUT_FILE" | \
        tail -n +5 | \
        grep -v -F ' --> ' | \
        grep -v '^$' 
)


printf "$TEXT\n"