A vinyl DJ companion built for crates, not clouds
Mobile app for vinyl DJs. Sync your Discogs collection, track BPMs, build playlists, run live sets. Offline-first, local storage, no subscriptions. Designed for dark booths, professionals, and collectors who rely on manual curation.
Vinyl DJs work offline, own physical objects, and dig through crates. Software must respect that.
I DJ with vinyl. Every time I prepared for a set, I juggled three disconnected tools: 1. Discogs for my collection metadata, 2. manual notes and stickies for tracking BPMs, and 3. my own memory for records and tracks. I asked myself many times: "Why is there not one tool that understands how vinyl DJs actually work?" Not streaming DJs scrolling through infinite playlists or amateurs without real crowd pressure. Vinyl collectors who perform live, often without internet, in poorly lit booths, surrounded by dozens of physical records they need to locate quickly.
The name came naturally: Kallax. After the ever-present IKEA shelving unit that has become the unofficial furniture of record collectors worldwide. Modular, unpretentious, functional. The objective was clear:
The tools available to vinyl DJs exist in isolated silos. Discogs serves as the ubiquitous repository for collection metadata, but offers no DJ functionality. Mainstream software solutions like Rekordbox are made to work with digital files, not physical objects. Streaming platforms become irrelevant the moment you are standing in front of a crate, and then there is the BPM problem... Unlike digital DJs who have tempo metadata embedded in their files, vinyl DJs need to know or feel what speed their tracks are. Some pay services to analyze tracks they do not even own digitally. Others manually tap out tempos on calculators, but most just guess and adjust by ear.
The opportunity was to design a mobile app that brings these disconnected pieces together. Sync with Discogs to import your collection. Fetch BPM data where APIs can provide it. Make manual entry feel natural rather than punitive. Offer a live mode that functions as passive reference during sets. No cloud lock-in, no monthly subscription, no illusion that perfect data exists for every record pressed in 1974. The constraint that shaped everything was to build it fast and alone, making every design choice serve the person standing in a dark booth trying to remember which crate holds their copy of that specific pressing.
Vinyl DJs and record collectors who catalog on Discogs, perform at venues, clubs, radio shows, private parties and value offline-first tools. We all appreciate when software understands the craft rather than trying to replace it.
Five principles emerged while designing Kallax. Each one responds to how vinyl DJs actually work—rather than how software typically imagines they should.
Honor the ritual. Digging through records is part of the craft. The muscle memory of flipping through crates matters.
Design for the booth. DJ booths are dark spaces, often with colored stage lighting. The interface had to work there.
Respect ownership. No login required. No cloud storage. No subscription model.
Accept reality. The GetSongBPM API returns tempo data for roughly 67% of tracks. So I built excellent manual entry instead of chasing perfect automation.
Stay out of the way. Three tabs, three concepts. No nested menus to hunt through under pressure.
Connect your Discogs account via OAuth and sync your complete collection: releases, tracks, artists, album art.
Two browse modes serve different needs. Grid view for efficient scanning. Cover-flow for the ritual of digging through crates one album at a time. Real-time search filters as you type.
Toggle between your full collection (SHELF) and favorited tracks (HOT). Tap any album to see the full tracklist with vinyl-specific positioning (A1, B2, etc.), track durations, and BPM badges.
After syncing, fetch BPM data from GetSongBPM API. A progress bar shows real-time status as it processes your library.
For tracks without data, or where the API got it wrong, tap the BPM badge to open a tap-tempo interface. Four taps gives you your tempo. Done.
Settings displays your BPM coverage percentage. No gamification. No pressure to hit 100%. Just honest reporting of what is verified versus what needs manual input.
Create playlists using tracks from your collection or hot tracks.
Drag-and-drop to reorder in manual mode. Or sort by BPM (ascending or descending) or alphabetically.
If a playlist exceeds 60 tracks, you will see a heavy bag warning. A practical reminder that carrying that many records to a gig is physically exhausting.
Playlists are local. They export with your backup. You own them completely.
Select a playlist, start a live session. The Now Playing card dominates the screen. Large album art, track title, artist, BPM clearly visible. Swipe left for next track, swipe right for previous. The queue scrolls below showing what is coming up. Skip to any track if you need to adjust on the fly. Edit BPMs mid-set if you discover the data was wrong. The screen stays awake automatically. Live Mode does not try to control your mixer or suggest what to play next. It assumes you know what you are doing and simply organizes the information you need to reference quickly.
Built kallax.app to match the app aesthetic. Single page, no trackers, no analytics, direct download links to both stores. The site had to communicate what the app does in under 10 seconds for visitors who found it through search.
First sketches in December 2025. First commit on Christmas Eve. Six weeks later it was live on both stores.
Building alone meant owning every decision: design, development, API integration, database, deployment, marketing, app store submissions. That kind of isolation forces clarity. If I could not explain why a feature mattered in one sentence, it did not matter enough to build.
So I cut. I cut a lot :) Collaborative playlists were too complex for the timeline. Spotify preview was nice-to-have, not essential. Genre and year filtering could wait until usage proved demand. What remained was the tightest possible loop: sync your collection, build a playlist, run a live set.
Solo building also meant accepting strategic imperfection. Six weeks of "good enough to be useful" beats "theoretically perfect" that never launches.
Kallax is now live on both the App Store and Google Play. In the weeks since launch, it has found its way onto 150+ devices. The measure of success for a tool like this is not download numbers or engagement metrics. It is trust. Do DJs trust Kallax when it matters? In the dark, without internet, in front of an audience? The professional DJs I know, friends who play clubs, radio shows, private events, are using it during live sets. No support questions about navigation or core functionality. Feature requests are refinements, not confusion about how things work. So far, yes.
Designing and building Kallax end-to-end taught me that constraints force honesty in product design. When you cannot hide behind "engineering will figure that out" or "we will improve that later". Every feature has to justify its existence immediately. The biggest shift in thinking for me, was designing for 67% coverage instead of 100% perfection. Most tools overpromise. Kallax does less, but what it does it does with transpaence and accuracy. That turned out to be more respectful to users than pretending certainty where none exists.
