cli script to sumarize a yt video

resume.sh

#!/usr/bin/env bash
set -euo pipefail

# summarize-youtube.sh
# Usage: ./summarize-youtube.sh "https://www.youtube.com/watch?v=A9BiNPf34Z4" [lang]
# Requires: yt-dlp, jq, llm
# Optional: set COOKIES_FROM_BROWSER=chrome (or your browser) if the video needs auth/age check.

if [[ $# -lt 1 ]]; then
  echo "Usage: $0 <youtube-url> [lang]" >&2
  exit 1
fi

for cmd in yt-dlp jq llm; do
  if ! command -v "$cmd" >/dev/null 2>&1; then
    echo "Missing dependency: $cmd" >&2
    exit 1
  fi
done

url="$1"
lang="${2:-en}"

# Optional cookies-from-browser support
extra_args=()
if [[ -n "${COOKIES_FROM_BROWSER:-}" ]]; then
  extra_args+=(--cookies-from-browser "$COOKIES_FROM_BROWSER")
fi

# Get video ID
id="$(yt-dlp --print id --skip-download "${extra_args[@]}" "$url")"

# Fetch captions (prefer human, fall back to auto), JSON3 format
yt-dlp -q --no-warnings --skip-download \
  --write-subs --write-auto-subs \
  --sub-langs "${lang}.*" \
  --sub-format json3 \
  -o '%(id)s' \
  "${extra_args[@]}" \
  "$url" >/dev/null 2>&1 || true

# Find the best matching captions file
capfile=""
if compgen -G "$id.${lang}"'*.json3' > /dev/null; then
  capfile="$(ls -1 "$id.${lang}"*.json3 | head -n1)"
elif compgen -G "$id."*'.json3' > /dev/null; then
  capfile="$(ls -1 "$id."*.json3 | head -n1)"
fi

if [[ -z "$capfile" ]]; then
  echo "No captions found (language: $lang). Try a different lang code or check --list-subs." >&2
  exit 2
fi

# Convert JSON3 captions -> plain text
out="/tmp/$id.txt"
jq -r '.events[]? | (.segs // []) | map(.utf8) | join("")' "$capfile" > "$out"

# Prompt + subtitles -> llm
{
  echo "<prompt>sumarize this video subtitle file and extract the gist of advice and learnings </prompt>"
  echo "<subtitles>"
  cat "$out"
  echo "</subtitles>"
} | llm