youtube-transcript-ingestionlisted

# YouTube Transcript Ingestion Batch-grab transcripts from any YouTube channel and save them as readable `.md` files -- with a guided 5-step flow and two rate-limit strategies. ## When to load this skill Load when the user wants to: - Research what a specific creator has said about a topic - Build a reading corpus from video content without watching every video - Filter content overload from a channel ("what here is actually new for me?") - Archive transcripts for future reference or AI-assisted analysis Trigger phrases: "grab YouTube transcripts", "ingest YouTube channel", "yt-dlp transcripts", "markitdown YouTube". --- ## What this skill does 1. Extracts the full video list from any YouTube channel (no download, metadata only) 2. Filters by date range + title keywords 3. Validates actual upload dates (flat-playlist timestamps are unreliable -- always verify) 4. Fetches full transcripts for each matched video (markitdown by default; yt-dlp + browser cookies when rate-limited) 5. Saves each as a `.md` file with clean YAML frontmatter into a folder of the user's choice 6. Optionally compares transcripts against an existing notes/context folder to surface only what's genuinely new --- ## Pre-flight check Before starting, verify the two required tools. Run this: ```bash echo "yt-dlp: $(which yt-dlp 2>/dev/null || echo NOT FOUND)" echo "markitdown: $(which markitdown 2>/dev/null || echo NOT FOUND)" # markitdown ships YouTube support as an OPTIONAL extra -- a bare i