video-transcript-downloader

video-transcript-downloader — video transcript extraction

Downloads auto-generated or manual subtitles and converts them to clean Markdown.

When to use

• "この YouTube 動画の文字起こしが欲しい"
• "この動画の内容をテキストにして"
• "transcript をダウンロードして要約したい"

Do not use for:

• Downloading audio or video files (use yt-dlp directly)
• Videos with no available subtitles (will fail gracefully)
• Paid/paywalled content

Extraction steps

# 1. Check available subtitles
yt-dlp --list-subs "<url>"

# 2. Download preferred subtitles (Japanese first, English fallback)
yt-dlp --write-subs --write-auto-subs --sub-lang "ja,en" \
  --skip-download --convert-subs srt \
  -o "%(title)s.%(ext)s" "<url>"

# 3. Convert SRT to clean Markdown (strip timestamps)
python3 - <<'EOF'
import re, sys
srt = open(sys.argv[1]).read()
lines = re.sub(r'\d+\n\d{2}:\d{2}:\d{2},\d{3} --> \d{2}:\d{2}:\d{2},\d{3}\n', '', srt)
lines = re.sub(r'\n{3,}', '\n\n', lines).strip()
print(lines)
EOF transcript.srt > transcript.md

Output

Clean Markdown file at ~/work/docs/transcripts/<video-title>.md or operator-specified path.

Hard-rule reminders

• Only download from publicly accessible URLs.
• Check license/terms before using transcript content commercially.
• yt-dlp must be installed: brew install yt-dlp.