yt-dlp

AI skill to analyze song requests, verify local workspace files, and download missing tracks directly from YouTube bypassing API limits.

yt-dlp (OpenClaw Music Fetcher)

Workflow: Analyze prompt -> Identify specific track -> Check ~/.openclaw/workspace/music/ folder -> Download if missing -> Pass to cmus.

Find the active workspace music folder

The agent must always operate within the specific OpenClaw workspace directory to ensure portability.

• Path: ~/.openclaw/workspace/music/ (Resolved as an absolute path from the user's home directory).
• Never download files to the root, current working directory, or system folders.

Track Identification & Verification Logic

When the user requests a song (via lyrics, artist, composer, or title):

1. Analyze: Determine the exact track name and artist.
2. Check Local: Search the workspace music folder: find ~/.openclaw/workspace/music/ -type f -iname "*<keyword>*"
3. Branching:
   • If the file EXISTS: Skip download. Pass the absolute file path to the cmus skill.
   • If the file is MISSING: Proceed to the download step using yt-dlp.

yt-dlp quick start

Ensure the directory exists, navigate to it, and download the missing track. The ytsearch1: prefix is used to grab the best match from YouTube Music/YouTube automatically:

• mkdir -p ~/.openclaw/workspace/music/ && cd ~/.openclaw/workspace/music/ && yt-dlp -x --audio-format mp3 -o "%(title)s.%(ext)s" "ytsearch1:<Exact Track Name> <Artist>"

Notes

• Post-Download Action: Always wait for the download to finish, then run find again to capture the final filename (as yt-dlp sanitizes titles) before passing it to cmus.
• Generalization: Use ~ or $HOME instead of hardcoded paths like /home/huy/ to ensure the skill works for all users.
• Dependency: yt-dlp requires ffmpeg to extract and convert audio streams into .mp3 format.
• Prefer highly specific search queries (Track + Artist) to ensure the agent picks the correct version of a song.