A song has more authors than any label prints. Sampld finds them, asks them, and makes sure they get paid.

Music rights are scattered across half a dozen databases that disagree with each other. Sampld is the missing plumbing: figure out who's actually on the hook, get their yes on their terms, and land the money in the right accounts.

Onboarding indie catalogues · Sync teams · AI platforms
SAMPLD ⁄⁄ LEDGER 038291 ⁄⁄ RESOLVED 2026·07·12 ⁄⁄ BLUE ROOM 4AM ⁄⁄ J.HOLLOW · R.VEGA · L.MARKS · D.FUNG ⁄⁄ SIDE A ⁄⁄ MADE IN LONDON ⁄⁄ SAMPLD · CLEARED Blue Room, 4AM JUNE HOLLOW LEDGER 038291 · 33⅓
Spinning / Blue Room, 4AM · June Hollow · Ledger 038291
01 · What broke

Sampling looks like one yes. It's usually seven, from people you can't find.

You need permission from two unrelated parties, and any one shareholder can kill the deal. On top of that, the public metadata is so bad the professional advice in 2026 is still "phone ASCAP and ask for the song indexing department."

a. Two rights

The recording and the song are owned by different people. Both have to agree.

The master usually belongs to a label. Sometimes the artist, sometimes whoever bought the catalogue thirty years ago. The song itself belongs to whoever wrote it — split into fractions between writers and their publishers. You need agreement from both sides, separately.

b. Fractional splits

A modern pop song has six credited writers. Every one holds a veto.

There's always someone. An estate that refuses on principle. A writer who doesn't like the new song's politics. Someone who quotes a number your project can't cover. One refusal ends it. You cut the sample, or you shelve the track.

c. No directory

There's nowhere central to look this up. Barely anywhere decentralised, either.

The PRO databases only cover the composition side. Finding a master after forty years of label mergers is basically detective work. Most producers give up and stick to royalty-free libraries, routing around the problem entirely.

02 · The stack

Three pieces, in order.

The order matters. Nothing works until people can find each other, so the registry comes first. Consent comes next, because it only makes sense once you know who to ask. Licensing sits on top of both, and only becomes useful when they're running.

01

Find every party who has to sign off on a recording.

Paste an ISRC or a Spotify link, drop a stem in, or just upload the audio. What comes back is the whole ownership graph: master owner, every writer, every publisher, whose PRO they're on, whose inbox actually reads the mail. Behind that, we're reconciling data from the PROs, the MLC, distributors, Discogs, label copy and direct registrations, and settling the conflicts when they disagree.

02

Let artists set their terms once, across everything they've made.

Rights holders publish a policy for their whole catalogue: what they'll allow (sampling, interpolation, stems, AI training, voice), at what rate, and where the carve-outs are. It's machine-readable and revocable at any time. When a producer sends a request, it either matches the policy and clears in a second, or it lands in the artist's inbox as an actual question. Nobody negotiates the same conversation twice.

03

Issue the license and register the splits properly the first time.

Once both sides agree, Sampld issues the license, sends the payment, and files the split on the new work. The point of doing this properly upfront is that the royalty flow is right from release day, instead of something someone has to reconstruct in a lawsuit three years in. Credits follow the work as it moves through streams, syncs, and whatever it gets sampled into next.

03 · Who it's for

Different people, same missing rail.

Artists & indie labels

Post one policy. Get inbound.

You decide what you'll allow, at what rate, and where the line is. When someone uses your work under those terms, the credit and the money actually show up — including for uses that until now had nowhere to land.

Producers

Clear it in a query.

Or find out you can't, in the same query. Either answer beats three months of unanswered email. Where the rights holder has posted their consent, the ask resolves against the rate on the spot.

Sync & clearance

Skip the manual resolution.

Route requests straight to the correct human. Batch-resolve a catalogue in one pass. Export a chain of title you can defend. Everything you already do by hand, done faster and with sources you can point to.

AI platforms

Prove provenance at scale.

An API for checking whether any input, output, or training set brushes against rights outside what you've already licensed. Built for teams with real licensing programs and settlements to keep clean.

04 · The opening move

Before anything we're selling, the lookup is free.

Producers, clearance lawyers, sync supervisors, music journalists. They all need to answer "who owns this and who do I ask" most weeks, and nobody has a decent tool for it. So that's the one we're giving away. Every correction makes the underlying graph better. Once Sampld is where people go to look this up, everything else we're building (consent, licensing, the AI provenance API) becomes possible.

05 · Get on the list

Early access is opening in waves through 2026.

We're onboarding indie distributors and labels first, then sync agencies and clearance firms, then AI platforms with active licensing programs. Tell us what you'd use it for; we'll route you into the right cohort.

· No spam. One email when your cohort opens.
· Your submission is not a promise, ours or yours.
· Written and answered by a human in London.
We keep this in a text file. Really.

You're on the list.

We'll be in touch when your cohort opens. In the meantime, keep making things.