Hi Spotify Team,
I've been using the Spotify MCP (Model Context Protocol) connector with Claude AI, and while I appreciate the integration, the current feature set feels quite limited compared to what the Spotify Web API actually offers.
Currently available via MCP:
- Search (tracks, artists, albums, playlists, podcasts)
- Create new playlists from natural language descriptions
- Read listening history and library
What's missing — and what would make this integration genuinely useful:
- Playback control — Play/pause, skip, set volume, play a specific track/album/artist/playlist on an active device. This is the most obvious gap. Being able to say "play Dark Side of the Moon on my living room speaker" is the core value proposition of an AI assistant integration.
- Playlist editing — Add tracks to existing playlists, remove tracks, reorder. Currently I can only create new playlists from a prompt, but cannot modify them or any others.
- Read playlist contents — Retrieve the actual track list of a playlist (mine or shared ones). Without this, the AI can see that a playlist exists but cannot work with its contents — making tasks like "copy my songs from this collaborative playlist" impossible.
- Library management — Save/unsave tracks, follow/unfollow artists, save albums.
- Queue management — Add tracks to the playback queue, view current queue.
Why this matters: Right now the MCP connector is essentially a search-and-recommendation tool with a "create playlist from vibe" feature on top. That's nice, but it doesn't take advantage of what makes AI assistants powerful: handling multi-step tasks across the user's actual Spotify data. The Web API already supports all of the above for Premium users — the limitation is purely in what the MCP server exposes.
Competing AI assistant integrations (and direct Web API access) offer these features. It would be great to see Spotify lead here rather than lag behind.
Thanks for considering!