youtube-scrape

Scrape YouTube videos: $ARGUMENTS

$ARGUMENTS should be one of:

• A YouTube channel handle (e.g. @AlanHirsch) or channel ID (e.g. UCxVxcTULO9cFU6SB9qVaisQ)
• --video VIDEO_ID to scrape a single video
• --refresh to re-fetch stats for all existing videos without re-downloading transcripts
• Empty — ask the user for the channel handle

Before Starting

1. Confirm YOUTUBE_API_KEY is set in .env.local — if not, tell the user to add it
2. Confirm yt-dlp is installed (which yt-dlp) — if not, tell the user to install it (pip install yt-dlp)
3. Read src/lib/database/schema.ts for the videos, videoSeries table definitions
4. Read scripts/youtube-scrape.ts to understand the existing scrape script

Pipeline Stages

Stage 1 — Fetch All Videos from Channel