lyricmood-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
lyricmood is the user's personal, lyrics-aware music assistant (Spotify, LRCLIB lyrics, Last.fm tags, ReccoBeats audio features).
The user is often not a native speaker of the song's language. When asked what a song means or why they might like it, call get_track_profile (or get_lyrics), then explain the themes in plain words, translate key lines, and explain idioms, slang and cultural references. Quote only short fragments of the lyrics.
"This song" / "what's playing" means now_playing. Refer to tracks by the
id returned from search_track. find_similar_by_lyrics can take about a
minute the first time for a song; say so before calling it. Tool results
include a missing field when a source had no data; mention it briefly
instead of guessing.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| logging | {} |
| prompts | {
"listChanged": true
} |
| resources | {
"subscribe": false,
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_artist_tracksA | An artist's most popular tracks (by Last.fm listeners), resolved to Spotify. Falls back to Spotify search order when Last.fm has no data. |
| analyze_artistsA | Sample each artist's most popular songs (lyrics excerpts, average audio mood, Last.fm tags) so you can find what they have in common, e.g. to answer "why do I like these artists?". Read the excerpts for recurring themes, imagery, perspective and emotional tone; use shared_tags and mood averages for the musical side. |
| get_my_top_tracksB | The user's most-played Spotify tracks over a time range, most played first. Good starting point for "why do I like what I like?". |
| get_saved_tracksA | The user's Liked Songs on Spotify, most recently liked first. |
| get_my_playlistsA | Playlists in the user's Spotify library (own and followed). Only
|
| get_playlist_tracksA | The songs of one of the user's own (or collaborative) playlists, in order. Needed before reordering or removing songs. |
| get_lyricsA | Get a song's plain lyrics (from LRCLIB) so you can translate, explain
or interpret them. Long lyrics are cut at a stanza boundary
( |
| get_track_profileA | Everything about one song in a single call: lyrics (possibly
truncated) with simple stats (repeated lines usually mark the chorus),
audio features from ReccoBeats (valence = musical positivity, energy,
danceability, acousticness, tempo, key), and Last.fm mood/genre tags.
Any source can be missing; |
| build_mood_playlistA | Pick and order songs for a mood, judging both what the lyrics are
about (mood_text) and how the music feels (valence/energy targets you
choose from the user's request). With |
| create_spotify_playlistA | Create a new playlist in the user's Spotify account (private unless
|
| update_spotify_playlistA | Change one of the user's own playlists: remove tracks, append tracks (ones already in it are skipped), and/or set a completely new order, in that sequence. This changes the user's account: confirm with the user first. |
| find_similar_by_lyricsA | Find songs whose lyrics are about similar things as the given song
(works across languages). Candidates come from Last.fm (similar tracks,
tags, similar artists) and ReccoBeats, and are ranked by lyrics meaning
with a local multilingual model. Scores are 0-1 relative to this search,
not absolute. |
| search_trackA | Search Spotify for tracks. Returns compact track info (id, title,
artists, album, year, duration, isrc). Use the returned |
| now_playingA | The song currently playing on the user's Spotify, with its profile (lyrics, audio features, tags) so you can explain it right away. Use it when the user says "this song", "what's playing", etc. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 14 tools
Most tools target distinct resource+action pairs, and the potentially confusing trio (get_lyrics vs get_track_profile vs now_playing) is explicitly disambiguated in the descriptions (plain lyrics vs full profile vs currently playing). The propose-vs-save split between build_mood_playlist and create_spotify_playlist is also clean. Only minor residual overlap remains, so an agent rarely misselects.
Names follow a consistent snake_case verb_noun pattern (get_*, create_*, update_*, build_*, find_*, search_*, analyze_*). The lone deviation is `now_playing`, which is a state noun rather than a verb_noun action. The convention is otherwise very predictable.
Fourteen tools sit comfortably in the well-scoped 3-15 range, and each one maps to a distinct capability (library reads, discovery, lyrics analysis, playlist writing). No tool feels redundant or padded.
The surface covers the core lifecycle: search/profile/lyrics, discovery via similar-by-lyrics and mood playlists, reading the user library, and creating/updating playlists. Gaps include no playlist deletion or unfollow and no dedicated album/artist-profile lookup, but agents can work around these for the stated lyric+mood purpose.