The latest articles on DEV Community by spotif (@spotify_ef33e87989). https://dev.to/spotify_ef33e87989 https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4218126%2F90a038e4-b58e-487e-b93f-40c9ca024bdd.png https://dev.to/spotify_ef33e87989 en spotif Sat, 31 Oct 2026 01:50:39 +0000 https://dev.to/spotify_ef33e87989/i-built-a-spotify-to-mp3-converter-heres-the-stack-40ie https://dev.to/spotify_ef33e87989/i-built-a-spotify-to-mp3-converter-heres-the-stack-40ie <p><em>A technical deep-dive into building a fast, free Spotify downloader that actually works.</em></p> <p>Every developer who's tried building a Spotify downloader hits the same wall: Spotify doesn't want you doing this. Their API doesn't expose audio files, and their streams are encrypted with Widevine DRM.</p> <p>Here's how I built <a href="https://spodown.com/en/" rel="noopener noreferrer">Spodown</a> — a free Spotify-to-MP3 converter that handles thousands of downloads daily.</p> <h2> The Architecture </h2> <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code>User paste Spotify link ↓ Backend resolves track metadata (name, artist, duration) ↓ YouTube Music search → find matching audio ↓ yt-dlp extracts audio stream ↓ ffmpeg transcodes to MP3/FLAC ↓ Serve file to user </code></pre> </div> <p>Simple in theory. The devil is in the details.</p> <h2> Stack </h2> <ul> <li> <strong>Frontend:</strong> Vue 3 + TypeScript + PWA</li> <li> <strong>Backend:</strong> Go + Fiber</li> <li> <strong>Matching:</strong> Custom YouTube Music search scoring</li> <li> <strong>Download:</strong> yt-dlp with residential proxy rotation</li> <li> <strong>Transcoding:</strong> ffmpeg</li> <li> <strong>Cache:</strong> MediaFire/VK document storage</li> <li> <strong>Queue:</strong> In-memory job system with SSE progress</li> </ul> <h2> The Matching Problem </h2> <p>The hardest part isn't downloading — it's finding the right audio. Spotify track metadata needs to match YouTube Music results, and the matching has to be <em>really</em> good or users get wrong songs.</p> <p>I built a scoring system that compares:</p> <ul> <li>Track name (exact &gt; partial &gt; fuzzy)</li> <li>Artist name</li> <li>Duration (±3 seconds tolerance)</li> <li>Album name (when available)</li> </ul> <p>Only candidates scoring above a threshold get accepted. Below that, we return "not found" rather than a wrong song.</p> <h2> The Widevine Path </h2> <p>For tracks that YouTube doesn't have (rare songs, new releases), I built a fallback: Widevine CDM decryption. This decrypts Spotify's own AAC streams directly — it's what the official Spotify Web Player does, just automated.</p> <h2> What I Learned </h2> <ol> <li> <strong>Caching is everything</strong> — 70% of downloads are repeat requests. Store results, skip the pipeline.</li> <li> <strong>Rate limits are real</strong> — both Spotify and YouTube throttle. Rotating proxies and account pools are essential.</li> <li> <strong>Users want simplicity</strong> — "paste link → get file" beats any feature list.</li> <li> <strong>SEO matters</strong> — technical tool ≠ invisible. I built the site server-rendered with proper meta tags so search engines could index it.</li> </ol> <h2> Try It </h2> <p><a href="https://spodown.com/en/" rel="noopener noreferrer">Spodown</a> — free tier needs no account. PRO unlocks 320kbps/FLAC if you want max quality.</p> <p><em>Questions? Drop them in the comments — happy to go deeper on any part.</em></p> spotify webdev opensource showdev