From "split a song" to a soul

Toneloom started with a simple, slightly naïve wish: what if you could pull any song apart and play with the pieces? Lift the vocal out of a favourite track. Mute the drums and sing the rest. Hear the bassline alone, the way the player heard it.

The first thing we learned is that the obvious version of that idea is impossible — and the second is that the impossible version was never the interesting one.

The wall we hit first

Everyone's instinct is "import from Spotify." It can't be done, and not for lack of trying. Streaming APIs hand you metadata and playback control — never the raw audio. Apple ties decoded frames to DRM; Spotify deprecated the analysis endpoints new apps could use. And even if you could grab the audio, separating and redistributing stems of a copyrighted master is a derivative work that needs a licence. "Credit to the original artist" is not a licence.

So we drew a hard line, the same one Moises drew before us: Toneloom only ever separates audio you bring in yourself — a file you own, an import you're entitled to. No rips. That constraint turned out to be a gift. It freed us to stop chasing other people's catalogues and start obsessing over the one thing we could make genuinely better: how the layers sound, and how it feels to play with them.

The playbook everyone uses — and why we tore it up

Open any AI audio tool and you'll see the same costume: near-black background, electric-blue waveform, neon-purple gradient, a font that whispers "engineer." It's competent. It's also cold, and it's everywhere. It makes music software feel like a spreadsheet that happens to make noise.

Music is warm. The tools to shape it had stopped being warm. That gap became our entire design brief.

We built Toneloom in a different language entirely — a folk-art one. Deep clay backgrounds instead of black. Turmeric, terracotta, indigo and peacock instead of neon. Serif headlines for warmth, clean sans for control. And at the centre of it all, a rangoli — the hand-drawn mandala you'd find chalked at a doorway — that forms as your song separates, breathes while it plays, and rests as a quiet emblem when you stop. It isn't decoration. It's the app's heartbeat.

Separation is the boring part

Here's the contrarian bet at the core of Toneloom: the split itself is a commodity. Good open models separate vocals and drums cleanly enough. If all you do is hand someone six dry stems, you've built a utility — and utilities compete on price until everyone loses.

The value is in what happens after the split. So we built the part nobody else bothers with: a room full of modelled analog gear — tube amps, turntables, tape machines, classic monitors — and the ability to run each layer through its own rig. Your vocal can glow through a warm tube while the drums hit a vintage console and the bass sits on tape. That per-layer rig is the thing you can't get anywhere else, and it's where a "song splitter" becomes an instrument.

The obsession that wouldn't quit: clean, clear, warm

Somewhere along the way "make it sound good" became a measurable discipline. We noticed a dampness in the low-mids that crept in with the warmth — pleasant for a second, muddy after a minute. So we built a Clarity control that de-muds and restores air without going clinical, and a deterministic "Listener" that scores every render on warmth, air and stage and refuses to ship something that's gone harsh or boomy.

Then we asked a harder question: can sound float? Can a vocal feel like it's hanging just in front of you, airy at the very top, while the band wraps the room? That became Acoustic Space — now on by default, tuned automatically to each song's key, tempo and mood. (We wrote up the physics of it here.)

Where we are

Toneloom runs on the web and on iOS and Android, with a backend that separates once and caches forever, and a design that's the same warm folk-art language on every surface — down to the marketing site you're reading this on. Next is a Pro cloud library that keeps your whole catalogue ready, Lightroom-style: separate once, and every rig you craft is saved as a non-destructive variant you can pull up in an instant.

We didn't set out to split songs. We set out to make a thing that loves music as much as you do, and is honest about what it can and can't do. The rangoli at the middle of the screen is the promise: this is made by hand, with care, for your ears.

Want to hear it? Get Toneloom on the App Store — twenty separations on us.