Swaddle: 30+ MCP Tools Connecting Claude AI to Spotify
Joey Musselman · December 30, 2025
I wanted to manage my Spotify library through conversation instead of clicking through menus. Describe what you want — "make me a workout playlist with high-energy tracks from my library" — and Claude handles the rest.
Challenge
Spotify's API requires OAuth 2.0 with automatic token refresh, rate limiting to avoid getting blocked, and intelligent caching to avoid re-fetching thousands of tracks on every query. The MCP server needed to sync a full music library (700+ tracks with 13 audio feature metrics each) to a local database for fast search, while exposing clean tools that Claude could use without hallucinating.
My Role & Contribution
Sole Developer — AI Tooling & API Integration
- Built 6 MCP tools: sync liked songs, search library, search Spotify catalog, create playlists, get sync status, and get library count
- Implemented full OAuth 2.0 flow with automatic token refresh and secure credential storage
- Designed PostgreSQL schema caching 13 audio features per track (danceability, energy, valence, tempo, etc.) for intelligent search
- Added rate limiting and batched API calls to stay within Spotify's limits
- This is one of 30+ production MCP tools I've built across multiple projects
Key Technologies
Impact
Fully functional MCP server that turns Claude Desktop into a Spotify power tool. Search, discover, and create playlists entirely through natural language. The caching layer means library searches are instant instead of waiting for API round-trips.