Completed PythonDesktop AppMusicAPI IntegrationVibe-Coded

Music Player — Spotify Alternative

A fully-featured desktop music player for local audio files with real-time synced lyrics, AI-powered recommendations, and custom playlist management.

GitHub ← All Projects

Overview

I created a complete music player that does what Spotify does, but for your local library. Real-time synced lyrics, genre-based discovery, personalized recommendations, and custom playlists — all running locally without paying a subscription.

This was built in close collaboration with ChatGPT, where I handled product direction and iterative refinement while ChatGPT provided the architecture and implementation.

Image of Functioning Application with Example Music

What I Built

The Problem

Spotify locks you into their catalog. Bandcamp albums you bought. Vinyl rips on your hard drive. Local files you created. All stuck in folders, unplayable in a real music app experience.

My Solution

A desktop music player that:

  • Plays local files from your computer
  • Syncs lyrics in real-time — highlights the current line as it plays
  • Recommends songs based on what you actually listen to
  • Auto-tags genres and lets you browse by category
  • Manages playlists — create, edit, reorder with ease

Key Features

Real-Time Synced Lyrics: I directed ChatGPT to implement lyrics downloading and line-by-line synchronisation. As the song plays, the lyrics highlight the current line in real time. It’s polished and feels premium.

Smart Recommendations: The app analyzes your listening history (which songs you play most, how many times) and suggests similar tracks from your library. It’s not a black box — it’s based on YOUR listening behavior.

Genre Auto-Tagging: Instead of manually tagging 500 songs, the app uses metadata from your audio files to automatically categorise music. You get instant genre-based browsing.

Custom Playlists: Create playlists, reorder them, see them update in real-time. Full CRUD operations without complexity.

Threading for Responsiveness: I insisted on background loading of metadata and lyrics. While you’re browsing, the app fetches data in the background — no frozen UI.

The Collaboration

My Role:

  • Defined the feature set and user experience
  • Decided what “premium” means (real-time lyrics, recommendations that work)
  • Iterated on the UI and functionality
  • Directed ChatGPT toward better approaches
  • Tested edge cases and requested refinements

ChatGPT’s Role:

  • Built the music playback engine
  • Implemented the lyrics fetching and synchronisation system
  • Designed the recommendation algorithm
  • Architected the threading system
  • Wrote the playlist management logic

This demonstrates strategic AI use — I didn’t build it alone, but I shaped every decision about what it should do and how it should feel.

Technical Highlights

Lyrics Synchronisation: The hardest part. Fetches lyrics from an API, parses timestamps, and highlights the current line as playback progresses. Requires precise timing synchronisation.

Recommendation Algorithm: Tracks your listening history (play counts, frequency) and suggests songs with similar metadata (artist, genre, era). Simple but effective.

Threaded Loading: Metadata and lyrics load on background threads. Playback never stalls.

Playlist Persistence: Playlists saved to JSON so they survive app restarts.

Why This Matters

This project shows:

  • Product vision — I could articulate what “great” looks like
  • AI collaboration — effective direction and iteration with an AI tool
  • UX thinking — features are designed for the user, not just technically possible
  • Full-featured application — not a prototype, a real, usable app
  • User-centric design — recommendations based on your data, not generic algorithms

Tech Stack

Python · threading · JSON · urllib · dataclasses · OS

Built in collaboration with ChatGPT. Product vision and direction by me, implementation by ChatGPT.