Spotify Skill

Control Spotify playback, search music, manage playlists, generate discovery playlists, and analyze listening habits via the Spotify Web API. Use when asked to play/pause/skip music, search for songs, manage playlists, get recommendations, check what's playing, or anything Spotify-related. Triggers on "spotify", "play music", "what song", "playlist", "spotify play", "spotify pause", "spotify search".

Control Spotify playback, manage your library, discover music, and analyze listening habits.

Prerequisites

1. Create a Spotify app at developer.spotify.com
2. Add a redirect URI (e.g. http://127.0.0.1:3000/callback)
3. Copy config.example.json to config.json and fill in your credentials
4. Required scopes are auto-requested during auth

bash

cp config.example.json config.json
# Edit config.json with your client_id and client_secret

First-Time Authorization

bash

python3 scripts/auth.py auth

Opens a browser for OAuth2. After authorizing, paste the code parameter from the redirect URL:

bash

python3 scripts/auth.py auth <code>

Tokens are saved to config.json and auto-refreshed. For headless setups, run auth on a machine with a browser, then copy config.json.

Commands

Run via python3 scripts/spotify.py <command> [args]

Playback

Command	Args	Description
now-playing	—	Current track, artist, album, progress
play	[uri]	Resume playback, optionally with a specific track URI
pause	—	Pause playback
next	—	Skip to next track
prev	—	Previous track
seek	position_ms	Seek to position in ms
volume	0-100	Set volume percentage
shuffle	on|off	Toggle shuffle
repeat	off|track|context	Set repeat mode
queue-add	uri	Add track to queue
queue	—	Get current queue
devices	—	List available devices
transfer	device_id	Transfer playback to device

Search & Discover

Command	Args	Description
search	query	Search tracks, artists, albums, playlists
recommendations	[track_id]	Get recommendations seeded from a track
audio-features	track_id	Danceability, energy, tempo, key, etc.

Library

Command	Args	Description
saved-tracks	[limit]	Get saved/liked tracks
save	id1 id2 ...	Save tracks to library
playlists	[limit]	List your playlists
playlist-tracks	playlist_id	Get tracks in a playlist
create-playlist	name [description]	Create a new playlist
edit-playlist	playlist_id [new_name]	Edit playlist name/description
add-to-playlist	playlist_id uri1 ...	Add tracks to playlist
remove-from-playlist	playlist_id uri1 ...	Remove tracks from playlist
play-playlist	playlist_id	Start playing a playlist from the beginning
smart-playlist	name [source] [limit] [time_range]	Auto-generate playlist (see below)
claw-list	[name] [seeds] [per_seed]	Discovery playlist (see below)

User Data

Command	Args	Description
me	—	Your Spotify profile
top-tracks	[short|medium|long_term]	Your top tracks
top-artists	[short|medium|long_term]	Your top artists
recent	[limit]	Recently played tracks

Smart Playlists

smart-playlist auto-generates playlists from your listening data:

bash

# From your top tracks (last 6 months)
python3 scripts/spotify.py smart-playlist "My Favorites" top 30 medium_term

# From your liked/saved tracks
python3 scripts/spotify.py smart-playlist "Liked Mix" saved 20

# Recommendations based on your top tracks (discovers new music)
python3 scripts/spotify.py smart-playlist "Discover Weekly DIY" top-recs 30 short_term

Sources: top (most played), saved (liked songs), top-recs (recommendations from top tracks) Time ranges: short_term (~4 weeks), medium_term (~6 months), long_term (years)

Claw-List (Discovery Playlists)

claw-list creates a discovery playlist with diverse seeds from your top/saved tracks, plus mood-matched related tracks:

bash

# Default: 5 seeds × 5 related tracks = 30 tracks
python3 scripts/spotify.py claw-list "My Mix"

# Custom: 8 seeds, 3 related each = 32 tracks
python3 scripts/spotify.py claw-list "Quick Mix" 8 3

# Larger: 10 seeds, 7 related each = 80 tracks
python3 scripts/spotify.py claw-list "Deep Dive" 10 7

How it works:

1. Picks diverse seed tracks from your top 50 + saved tracks (deduplicates by artist for variety)
2. For each seed: analyzes audio features (energy, valence, danceability, etc.) to determine mood
3. Adds 2 same-artist deep cuts + (per_seed - 2) mood-matched discovery tracks via search
4. Creates the playlist with all tracks interleaved (seed, related, seed, related...)

Agent Usage Patterns

1. "Play [song name]" → search the song, then play the first result's URI
2. "What's playing?" → now-playing
3. "Skip/pause/next" → direct command
4. "Make me a playlist of..." → search for tracks, create-playlist, add-to-playlist
5. "Music like X" → recommendations seeded from X's track ID
6. "My top artists this month" → top-artists short_term
7. "Make me a discovery mix" → claw-list with desired size

URIs

Spotify uses URIs like spotify:track:4cOdK2wGLETKBW3PvgPWqT. Extract the ID from URLs like https://open.spotify.com/track/4cOdK2wGLETKBW3PvgPWqT — the ID is the part after /track/.

API Migration Notes (February 2026)

Spotify's February 2026 API migration removed several endpoints for dev-mode apps. This skill handles them:

• Playlist tracks: Uses /playlists/{id}/items (new) instead of /tracks (removed). Response field changed from track to item.
• Create playlist: Uses /me/playlists instead of /users/{id}/playlists (removed).
• Recommendations: /recommendations and /related-artists may return 404/403 in dev mode. Claw-list uses search-based discovery as a fallback.
• Batch endpoints: /tracks, /artists, /albums, /browse/new-releases are removed in dev mode.

Apps in Extended Quota Mode are unaffected. Request extended quota at the Spotify Developer Dashboard if needed.

Config

• config.json stores OAuth2 tokens and credentials — do not commit or share
• Tokens auto-refresh via auth.py (called transparently by spotify.py)
• config.example.json is the template for new setups
• Env vars SPOTIFY_CLIENT_ID, SPOTIFY_CLIENT_SECRET, SPOTIFY_REDIRECT_URI override config.json if set

Requirements

• Python 3.8+
• A Spotify Premium account (for playback controls)
• No external Python packages needed — uses only stdlib