song-identifier - SKILL.md Agent Skill

name: song-identifier description: Use when a user sends an audio file and asks "what is this song?", "identify this track", "shazam this", or similar. Accepts any audio format (OGG, MP3, WAV, M4A). Returns title, artist, album, release date, and Spotify/Apple Music links via the AudD API. metadata: emoji: 🎵 category: media platform: any

Song Identifier

Identify songs from audio files via the AudD music recognition API. Zero dependencies — uses Python stdlib only.

Requirements

Python 3.8+
AudD API token — free tier at https://audd.io/
Internet connection

Setup

export AUDD_API_TOKEN="your_token_here"

Add to ~/.zshrc or ~/.bashrc to persist.

Usage

python3 scripts/identify.py <audio_file_path>

Example:

python3 scripts/identify.py /tmp/clip.ogg

Output

🎵 Song Title
👤 Artist Name
💿 Album Name
📅 2023-05-12
🟢 Spotify: https://open.spotify.com/track/...
🍎 Apple Music: https://music.apple.com/...

Integration

Wire into an agent skill dispatcher for WhatsApp triggers like "what's this song?":

import subprocess
result = subprocess.run(
    ["python3", "scripts/identify.py", audio_path],
    capture_output=True, text=True,
    env={**os.environ, "AUDD_API_TOKEN": "..."}
)
return result.stdout

Scripts

scripts/identify.py — Main script. stdlib-only, no requests dependency. Reads AUDD_API_TOKEN from env.