gettranscript.sh

#!/bin/bash
# USAGE
# gettranscript.sh url [lang]
set -euo pipefail
URL="$1"
LANG="${2:-en}"
p=`mktemp`
dir=$(dirname "$p")
file=$(basename "$p")
cd "$dir"
yt-dlp \
	-o "$file" \
	--write-subs \
	--write-auto-subs \
	--sub-langs "$LANG" \
	--skip-download \
	"$URL" > /dev/stderr
ffmpeg -i "${file}.${LANG}.vtt" "${file}.${LANG}.srt"
cat "${file}.${LANG}.srt" | grep -v -- "-->" | grep -v -P '^[0-9]+' | grep -v -P -- '^\s*$' | dos2unix | uniq