🎵 Workshop Music MCP Server

AI @Princeton x CBC Workshop - October 30, 2025

Build a real MCP server that connects Claude to Spotify! Search songs, analyze music collections, and create playlists using natural language.

🎯 What You'll Learn

• MCP Fundamentals: Build servers that give Claude new capabilities

• API Integration: Connect to real services (Spotify)

• Tool Design: Create intuitive interfaces for AI

• Real-world Skills: Handle API changes and deprecations (like we did with Spotify's 2025 update!)

⚡ Quick Start

Prerequisites

• Python 3.10 or higher

• A Spotify account (free is fine!)

• Claude Desktop app installed

• Basic familiarity with terminal/command line

Installation

1. Clone this repository

   git clone <your-repo-url> cd workshop-music-mcp

2. Install dependencies

   pip install -r requirements.txt

3. Set up Spotify credentials

   • Follow SPOTIFY_SETUP.md to get your API keys

   • Copy .env.example to .env and fill in your credentials

4. Configure Claude Desktop

   • See SETUP.md for detailed instructions

   • Add the MCP server to your Claude Desktop config

5. Start the server and try it!

   python music_server_updated_2025.py

   Then restart Claude Desktop and start chatting!

🎵 What Can It Do?

Once connected, you can ask Claude things like:

Basic Discovery:

• "Find me some upbeat indie songs" - Search tracks

• "Tell me about Taylor Swift's most popular songs" - Artist information

• "Give me recommendations based on Blinding Lights" - Get song recommendations

Collection Analysis:

• "Is this playlist family-friendly?" - Check explicit content

• "How diverse is my music taste?" - Artist, genre, and era diversity

• "Which artists do I listen to most?" - Artist frequency analysis

• "What genres dominate my collection?" - Genre breakdown and distribution

• "Analyze this playlist" - Get popularity stats and explicit content overview

Playlist Creation: 🎯

• "Create a playlist from these 20 songs" - Make actual Spotify playlists

• "Balance this collection by genre and create a playlist" - Smart playlist generation

🔧 Available Tools (10 Total)

Discovery & Search

search_tracks

Search for songs on Spotify by name, artist, or keywords. Returns track details, popularity, and Spotify links.

get_recommendations

Get song recommendations based on seed tracks, artists, or genres. Great for discovering new music similar to what you already love.

get_artist_info

Get detailed information about artists including their top tracks, genres, popularity, and follower count.

Playlist Tools ⭐ NEW!

create_playlist

Create a real Spotify playlist from a collection of songs! Provide song names and artists, and it'll search for them and create the playlist on your account.

generate_balanced_playlist

Create a balanced playlist from a larger collection. Balance by:

• Genre - Equal representation across genres

• Artist - Max 2 songs per artist for diversity

• Era - Even distribution across decades

analyze_playlist

Analyze any Spotify playlist to get popularity stats, explicit content percentage, and track details.

Collection Analysis Tools 🎯

analyze_explicitness

Check explicit content in your song collection - find out what percentage has explicit lyrics and get a family-friendly rating.

analyze_collection_diversity

Analyze how diverse your music collection is:

• Artist variety and diversity score

• Genre spread and unique genres

• Popularity distribution (mainstream vs underground)

• Era range and distribution across decades

get_top_artists_from_collection

Find which artists appear most frequently in your collection and see their contribution percentages. Perfect for understanding your listening habits.

analyze_genres_in_collection

Get a complete genre breakdown - dominant styles, genre diversity score, and detailed distribution percentages with genre tags for each song.

🚨 What Changed in 2025?

Important API Update: Spotify deprecated their audio-features endpoint in November 2024. This MCP server has been updated to work without it!

What's different:

• ❌ Removed: Audio features like danceability, energy, valence, tempo are no longer accessible

• ✅ Still Available: Track info (name, artist, album, popularity, explicit flag), artist genres, release dates

• 🎯 New Focus: Playlist creation, genre analysis, explicitness checking, and collection diversity

What this means for you:

• You won't see "mood" analysis or "energy levels" anymore

• Focus shifted to practical tools: creating playlists and analyzing your collection's diversity

• The tools now use what's reliably available: popularity scores, genres, explicit flags, and release years

🏗️ Architecture

How it works:

1. You ask Claude a music question in natural language

2. Claude decides which MCP tool to use

3. Your MCP server calls Spotify's API

4. Results come back to Claude

5. Claude presents them in a friendly way

📚 Documentation

• QUICKSTART.md - Get running in 10 minutes

• SETUP.md - Detailed installation guide

• SPOTIFY_SETUP.md - How to get API credentials

• TOOLS_REFERENCE.md - 📄 Quick reference for all tools (print this!)

• EXAMPLES.md - ⭐ Example queries for bulk analysis tools!

• ADVANCED_ANALYSIS.md - 🎯 Advanced analysis: explicitness, diversity, artists, genres!

• WORKSHOP_GUIDE.md - Full workshop walkthrough

• TROUBLESHOOTING.md - Common issues and fixes

• CONTRIBUTING.md - Build your own tools!

🎓 Learning Resources

New to MCP?

• Anthropic MCP Documentation

• MCP Specification

Want to Learn More?

• Spotify API Reference

• Spotipy Documentation

• Python Async Programming

🏆 48-Hour Challenge

After the workshop, try building your own MCP server! Some ideas:

• 📚 Book Recommendations - Connect to Goodreads or Open Library

• 🎬 Movie Analysis - Use TMDB API for film data

• 🍕 Recipe Finder - Integrate with Spoonacular

• ⚡ Smart Home - Control your lights or thermostat

• 📊 Data Viz - Generate charts from your data

Join the AI@Princeton Discord for support during the challenge!

🐛 Troubleshooting

"ModuleNotFoundError: No module named 'spotipy'"

• Run: pip install -r requirements.txt

"Can't connect to Spotify"

• Check your .env file has the correct credentials

• Make sure you've created a Spotify app on their dashboard

"Claude can't see the tools"

• Restart Claude Desktop after adding the server to config

• Check that the server path in config is absolute (full path)

More issues? See TROUBLESHOOTING.md

🤝 Contributing

Want to add more features? Check out CONTRIBUTING.md for:

• How to add new tools

• Code style guidelines

• Testing your changes

• Submitting improvements

📝 License

MIT License - see LICENSE file for details.

🙏 Acknowledgments

Built for AI@Princeton's MCP Workshop Series. Special thanks to:

• Anthropic for creating MCP

• Spotify for their excellent API

• The Princeton AI community

📧 Questions?

• Workshop Questions: Ask during the session or on Discord

• Technical Issues: Open an issue on GitHub

• General MCP Help: Check docs.anthropic.com/mcp

Ready to build? Start with QUICKSTART.md or jump into SETUP.md!

⭐ If this helped you, give it a star on GitHub!