youtube

name: youtube description: Work with YouTube videos from the terminal using yt-dlp — fetch metadata, download audio/video, and pull transcripts/subtitles to summarize. Keyless. Use when asked to "summarize this YouTube video", "download audio from", "get the transcript of", or given a YouTube link. allowed-tools: ["@coder", "Bash", "@webfetch", "@read"] triggers: - youtube - summarize this video - download audio from - get the transcript - transcript of - resumir esse vídeo - baixar áudio de - transcrição do vídeo

Keyless YouTube access via yt-dlp (no API key). Detect: command -v yt-dlp ffmpeg (Unix) / Get-Command yt-dlp -ErrorAction SilentlyContinue (Windows). Install: brew install yt-dlp ffmpeg / pipx install yt-dlp / choco install yt-dlp.

Metadata (no download)

yt-dlp --dump-json --no-warnings "URL" | python3 -c "import sys,json;d=json.load(sys.stdin);print(d['title'],'-',d['uploader'],'-',d['duration_string'])"

Transcript / subtitles (best for summarizing)

yt-dlp --skip-download --write-auto-subs --sub-langs "en,pt" --sub-format vtt -o "%(id)s" "URL"

This writes a .vtt; read it with @read, strip timestamps, and summarize. Prefer manual subs (--write-subs) when available, auto-subs (--write-auto-subs) otherwise.

Download

yt-dlp -f "bestaudio" -x --audio-format mp3 "URL"     # audio only
yt-dlp -f "bv*+ba/b" -o "%(title)s.%(ext)s" "URL"      # best video+audio

"Summarize this video" workflow

Get metadata (title/author/length).
Pull the transcript (subs). If none exist, say so — don't fabricate.
Read the transcript and produce a structured summary (key points + timestamps if useful).

Rules

Respect copyright/ToS: download only what the user is entitled to; default to transcript-only for summarization.
ffmpeg is required for audio extraction/merging — check it in Step 0.