Velvet
"I don't rush. Neither should you."
Disco is a DJ you can talk to. Real songs by the real artists. Bespoke tracks made up on the spot. And a voice in your ear that actually reads the room.
No card. No playlist to build. Just talk.
Somewhere between the radio and the algorithm, we lost the person. Disco is the person coming back — with everything the algorithm learned still in their pocket.
A voice at 2am who knew what you needed before you did. Introduced you to your favourite band. Said your name on air. Made you wait for the good one.
Everything ever recorded, instantly, for the price of a sandwich. And a silent machine that watches what you skip and quietly narrows the world to your own reflection.
A voice again. It has the whole library, it knows your Sunday, it can invent a song that doesn't exist yet — and you can just tell it, out loud, like a human being.
You've been your own DJ since 2008.
Take the night off.
No search bar. No queue to babysit. You say the thing you'd say to a friend with good taste, and the room changes.
Not five filters with names on them. Five different people, with five different opinions about what you should be listening to — and five different amounts they'll say about it.
"I don't rush. Neither should you."
"Ten seconds. Trust me."
"The sun's out somewhere. We'll go there."
"Everything came from somewhere. Let me show you."
"Fewer words. Longer nights."
More residents land every season — and on Resident you can train your own, voice and taste, and keep them private.
A fully licensed catalogue. When you ask for the song, you get the record — the actual master, the actual voice. Not a soundalike, not a cover.
"Play the one she does live where the band drops out."
Describe it and it's written, performed and streaming in seconds — not rendered later, not queued. It starts playing while it's still being made.
"A lullaby for a baby called Ada. Put her mum's name in the second verse."
Two genres, two decades, two cities. It finds the actual overlap and builds a set that lives there — then tells you what it did.
"Bossa nova × drill. Keep it under 90 BPM."
Voices in the Artist Vault are licensed by the artists themselves, who get paid every time you use them. Consent first — always.
"What if she sang over that beat instead?"
Dedications, countdowns, birthday shout-outs, the "this next one is for" moment. Ducked under the track, timed to the bar, in your DJ's own voice.
"It's my mum's 60th. Say something at midnight."
Beatmatched, key-aware transitions that hold a room. Nothing ever just stops. The set has an arc, and it knows where you are in it.
"Don't stop for the next two hours."
Everyone in the room scans in from their own phone. Requests come to the DJ, not to you. It weighs the room, spaces out the wildcards, lets the birthday person cut the line, and quietly buries the one your friend keeps asking for. You never have to be the bad guy again.
"Twelve of us, mixed ages, nobody's mad by midnight."
And your gym, and your drive, and the album you always cry to. Over a few weeks it stops asking and just starts correctly.
"The usual."
One button does almost everything. The rest of the screen is there for when you want to look at what's happening — not because you have to.
Disco doesn't distinguish in the moment — it just plays. But everything it invents gets its own cover, its own credits, and a permanent home in your library.
A generated duet is only magic if the artist is in on it. So the Vault is opt-in, revocable, and metered — artists set what their voice can and can't be used for, and they're paid per generation, not per quarter.
Nothing outside the Vault is ever voice-modelled. Bespoke tracks by default use Disco's own session singers, who are paid and credited too.
Artists — get in touchGenres, moods, languages, contexts. No political use, no brand use, no explicit — whatever your line is, it's enforced before generation, not after.
What was generated, for whom, from what prompt. You can hear it. You can pull it. Your dashboard is the same data ours is.
Not per stream, per creation — plus streams after that. A collab someone plays a hundred times pays a hundred times.
Revoke and the voice is out of the model's reach within 24 hours. Already-generated tracks stop distributing. No perpetuity clause, ever.
The free tier isn't a demo. It's the whole conversation, with ads between sets and one resident on rotation.
Artists are paid on every tier — including the free one. Ads fund it, not the other way round.
The apps are in the last stretch of build. Pick your platform and you'll get the download the day it ships — plus a month of Backstage, on us, for being early.