yt-dlp

yt-dlp

Overview

Download and extract media from YouTube and 1000+ other sites using yt-dlp — the actively maintained fork of youtube-dl. This skill covers audio extraction (MP3/FLAC/WAV), video download with quality selection, playlist handling, subtitle download, metadata extraction, thumbnail embedding, batch operations, channel archiving, and integration with processing pipelines (ffmpeg, whisper).

Instructions

Step 1: Installation

# pip (recommended)
pip install yt-dlp

# Standalone binary
curl -L https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp -o /usr/local/bin/yt-dlp
chmod +x /usr/local/bin/yt-dlp

# Update
yt-dlp -U

# Verify
yt-dlp --version

ffmpeg is required for merging formats and audio conversion:

apt install -y ffmpeg    # Ubuntu/Debian
brew install ffmpeg      # macOS

Step 2: Audio Extraction

# Download best audio, convert to MP3
yt-dlp -x --audio-format mp3 "https://youtube.com/watch?v=VIDEO_ID"

# Best audio as FLAC (lossless)
yt-dlp -x --audio-format flac "URL"

# Best audio as WAV
yt-dlp -x --audio-format wav "URL"

# MP3 with specific quality (0=best, 9=worst)
yt-dlp -x --audio-format mp3 --audio-quality 0 "URL"

# Keep original format (no conversion)
yt-dlp -x "URL"

# Download audio + embed thumbnail as album art
yt-dlp -x --audio-format mp3 --embed-thumbnail "URL"

# Download audio + embed metadata (title, artist, etc.)
yt-dlp -x --audio-format mp3 --embed-metadata --embed-thumbnail "URL"

# Custom output filename
yt-dlp -x --audio-format mp3 -o "%(title)s.%(ext)s" "URL"
yt-dlp -x --audio-format mp3 -o "%(uploader)s - %(title)s.%(ext)s" "URL"

Step 3: Video Download

# Best quality (video + audio merged)
yt-dlp "URL"

# List available formats
yt-dlp -F "URL"
# Output: ID  EXT  RESOLUTION  FPS  FILESIZE  CODEC  BITRATE

# Download specific format by ID
yt-dlp -f 137+140 "URL"    # 1080p video + best audio

# Best video up to 1080p
yt-dlp -f "bestvideo[height<=1080]+bestaudio/best[height<=1080]" "URL"

# Best video up to 720p, prefer MP4
yt-dlp -f "bestvideo[height<=720][ext=mp4]+bestaudio[ext=m4a]/best[height<=720]" "URL"

# Only 4K if available
yt-dlp -f "bestvideo[height>=2160]+bestaudio" "URL"

# Smallest file
yt-dlp -f "worstvideo+worstaudio/worst" "URL"

# MP4 output (re-mux if needed)
yt-dlp --merge-output-format mp4 "URL"

Step 4: Playlists & Channels

# Download entire playlist
yt-dlp "https://youtube.com/playlist?list=PLAYLIST_ID"

# Playlist: audio only
yt-dlp -x --audio-format mp3 "PLAYLIST_URL"

# Download specific items from playlist
yt-dlp --playlist-start 5 --playlist-end 10 "PLAYLIST_URL"    # Items 5-10
yt-dlp --playlist-items 1,3,5,7-10 "PLAYLIST_URL"             # Specific items

# Download entire channel
yt-dlp "https://youtube.com/@ChannelName/videos"

# Download channel with organized folders
yt-dlp -o "%(uploader)s/%(playlist)s/%(title)s.%(ext)s" "CHANNEL_URL"

# Only videos from last 30 days
yt-dlp --dateafter today-30days "CHANNEL_URL"

# Skip already downloaded (archive file)
yt-dlp --download-archive archive.txt "PLAYLIST_URL"
# Subsequent runs skip already downloaded videos

Step 5: Subtitles

# Download video + subtitles
yt-dlp --write-sub --sub-lang en "URL"

# Download auto-generated subtitles
yt-dlp --write-auto-sub --sub-lang en "URL"

# All available subtitles
yt-dlp --write-sub --all-subs "URL"

# Subtitles only (no video)
yt-dlp --skip-download --write-sub --sub-lang en "URL"

# Convert subtitles to SRT
yt-dlp --write-sub --sub-lang en --convert-subs srt "URL"

# Embed subtitles into video file
yt-dlp --embed-subs --sub-lang en "URL"

# List available subtitle languages
yt-dlp --list-subs "URL"

Step 6: Metadata & Information

# Print video info (no download)
yt-dlp --dump-json "URL" | python3 -m json.tool

# Extract specific fields
yt-dlp --print "%(title)s | %(duration)s | %(view_count)s" "URL"

# Get thumbnail URL
yt-dlp --get-thumbnail "URL"

# Download thumbnail only
yt-dlp --skip-download --write-thumbnail "URL"

# Download description
yt-dlp --skip-download --write-description "URL"

# Download comments
yt-dlp --skip-download --write-comments "URL"

# Get video info as JSON for a playlist
yt-dlp --dump-json --flat-playlist "PLAYLIST_URL"

Step 7: Output Templates

# Fields: title, uploader, upload_date, duration, view_count, ext, id, playlist_index, etc.
yt-dlp -o "%(uploader)s/%(upload_date)s - %(title)s.%(ext)s" "URL"
yt-dlp -x --audio-format mp3 -o "podcasts/%(playlist)s/%(playlist_index)03d - %(title)s.%(ext)s" "PLAYLIST_URL"
yt-dlp -o "%(title).100B.%(ext)s" "URL"    # Limit title to 100 bytes

Step 8: Batch Operations

# Download from URL list, with archive tracking and rate limiting
yt-dlp -a urls.txt -x --audio-format mp3
yt-dlp -a urls.txt --download-archive done.txt -x --audio-format mp3
yt-dlp --rate-limit 5M --sleep-interval 5 -a urls.txt
yt-dlp -N 4 "PLAYLIST_URL"    # 4 concurrent downloads

Step 9: Other Platforms

yt-dlp supports 1000+ sites beyond YouTube. Use yt-dlp --list-extractors to see all:

yt-dlp "https://twitter.com/user/status/123456789"           # Twitter/X
yt-dlp "https://www.instagram.com/p/POST_ID/"                # Instagram
yt-dlp "https://www.tiktok.com/@user/video/123456789"        # TikTok
yt-dlp "https://www.twitch.tv/videos/123456789"              # Twitch VODs
yt-dlp "https://soundcloud.com/artist/track-name"            # SoundCloud
yt-dlp "https://vimeo.com/123456789"                         # Vimeo

Step 10: Pipeline Integration

# Download → extract audio → transcribe with Whisper
yt-dlp -x --audio-format wav -o "temp.%(ext)s" "URL"
whisper temp.wav --model small --output_format srt

# Download podcast → normalize → generate waveform
yt-dlp -x --audio-format wav -o "episode.%(ext)s" "URL"
sox episode.wav normalized.wav norm -1 highpass 80
audiowaveform -i normalized.wav -o waveform.json --pixels-per-second 20

Step 11: Configuration File

Save defaults in ~/.config/yt-dlp/config:

-f bestvideo[height<=1080][ext=mp4]+bestaudio[ext=m4a]/best[height<=1080]
-o %(uploader)s/%(title)s.%(ext)s
--embed-metadata
--embed-thumbnail
--download-archive ~/.local/share/yt-dlp/archive.txt
--rate-limit 10M
--sleep-interval 3
--write-auto-sub
--sub-lang en
--convert-subs srt

Examples

Example 1: Download a YouTube playlist as MP3 with metadata and thumbnails

User prompt: "Download all videos from this YouTube playlist as high-quality MP3 files with embedded album art and metadata. Organize them by playlist index number. Playlist URL: https://youtube.com/playlist?list=PLrAXtmErZgOeiKm4sgNOknGvNjby9efdf"

The agent will:

1. Verify yt-dlp and ffmpeg are installed.
2. Run yt-dlp -x --audio-format mp3 --audio-quality 0 --embed-thumbnail --embed-metadata -o "%(playlist_index)03d - %(title)s.%(ext)s" --download-archive archive.txt "PLAYLIST_URL".
3. The archive file ensures re-running the command skips already-downloaded tracks.
4. Report how many files were downloaded and their total size.

Example 2: Extract audio from a conference talk and generate subtitles

User prompt: "Download this conference talk as 720p MP4 with English subtitles embedded, then also extract just the audio as WAV for transcription: https://youtube.com/watch?v=dQw4w9WgXcQ"

The agent will:

1. Download the video with embedded subtitles: yt-dlp -f "bestvideo[height<=720]+bestaudio" --merge-output-format mp4 --embed-subs --sub-lang en --write-auto-sub "URL".
2. Extract audio separately: yt-dlp -x --audio-format wav -o "talk-audio.%(ext)s" "URL".
3. Confirm both files exist and report their sizes and durations.

Guidelines

• Always install ffmpeg alongside yt-dlp; it is required for merging separate video and audio streams and for audio format conversion.
• Use --download-archive archive.txt when downloading playlists or channels to avoid re-downloading videos on subsequent runs.
• Apply rate limiting with --rate-limit 5M --sleep-interval 5 when batch downloading to avoid being throttled or blocked by the source platform.
• Use -f "bestvideo[height<=1080]+bestaudio/best[height<=1080]" as a default format selector to get good quality without unnecessarily large 4K files.
• Keep yt-dlp updated regularly with yt-dlp -U; extractors break frequently as platforms change their APIs and page structures.