The ISRC Code, Explained for People Actually Releasing Music

An ISRC code is the twelve-character identifier that tells every royalty system, chart compiler and content-matching database in the world which specific recording it is looking at. It stands for International Standard Recording Code, it is defined by ISO 3901, and it is administered internationally by IFPI through a network of national agencies. If you have ever released a track through a distributor, you already have one, whether or not you noticed.

Most explanations stop there and tell you to let your distributor handle it. That advice is correct for the majority of releases, which is precisely why nobody explains the parts where it stops being correct — what the four segments mean, which of the two routes to take, and the version rules deciding whether a track gets a new code or reuses the old one.

This page is the identifier reference the rest of this site points at: structure, acquisition, versioning, and what the code does downstream — plus the one thing it emphatically does not do, which is tell anyone anything about how the recording was made.

An ISRC identifies a recording, not a song

This is the distinction everything else rests on. Three identifiers circulate in music admin and they refer to three different objects.

Read across those rows and the edge cases resolve themselves. A cover version of your song shares its ISWC and gets its own ISRC — same composition, different recording. A remaster shares the ISWC and needs a new ISRC, because the master changed. And a track released first as a single and later on an album keeps one ISRC across both while sitting inside two different UPCs.

The ISWC side is administered by your PRO rather than your distributor, which is a genuinely separate registration with a separate organisation. If you have not looked at that half of the picture, our ASCAP versus BMI comparison covers the societies; everything below concerns the recording side only.

Reading the code, character by character

An ISRC is twelve alphanumeric characters in four segments. People write it with hyphens for legibility — the hyphens are display formatting and are not part of the code as it is stored or transmitted.

Take GB-A1B-26-00001, stored as GBA1B2600001:

Three things about that table trip people up.

The country code describes the registrant, not the music. It is not where the track was recorded, where the artist lives, or where the release is sold. If your distributor's registrant code was issued in a different country from yours, your ISRCs will carry that country's prefix, and nothing is wrong.

You will also see codes that begin with Q. Q-prefixes are allocated to registrants that are not tied to a single national agency, which is why a large share of distributor-assigned ISRCs start with something like QM or QZ rather than a country you recognise. If you were told to sanity-check an ISRC by matching the country code to your distributor's location, that heuristic breaks here — and breaking it is the common case, not the exception.

The year is the year of assignment. A recording made in 2023 and assigned an ISRC in 2026 carries 26. The designation code runs from 00001 to 99999 and resets each year, which gives each registrant just under 100,000 assignments per calendar year — comfortable for a label, and not something an independent artist will ever approach.

The two routes to getting one

There are exactly two ways an ISRC comes into existence for your recording: someone else assigns it, or you do.

Route one — your distributor assigns it

Upload a track to DistroKid, TuneCore, CD Baby or any of the alternatives in our music distribution services comparison and an ISRC is generated for each track automatically, free, with no application and no forms. The codes are included in the metadata package delivered to streaming platforms and appear in your distributor dashboard afterwards.

This is how the overwhelming majority of independent artists get their codes — the figure that circulates is upwards of 95%, and it looks about right. It is free, it is instant, and it works. The distributor is acting as what the US agency calls an ISRC manager: a registered party assigning codes on behalf of clients. SoundExchange notes that ISRC managers are most often distributors, but can also be mastering engineers, labels or anyone else assigning on a rightsholder's behalf.

Route two — apply for your own registrant code

Apply to your national ISRC agency and you get your own three-character registrant code, after which you assign ISRCs yourself. In the United States that agency is the RIAA; in the United Kingdom it is PPL; in India it is PPL India. IFPI also runs an online registration route, priced in the region of £15 for a block of 30 assignment credits, for registrants without a more convenient national path.

What you get is control. What you take on is a workflow: a catalogue log, a naming convention, and the discipline to record every assignment before you need it back. The best operational advice we found on this, from an artist-services perspective, is to build the spreadsheet before you request the codes. Self-assignment fails when someone gets registrant access first and then tries to invent the catalogue logic afterwards.

Which one you should actually pick

Self-assignment starts to pay when you have a growing catalogue and want one internal source of truth, when you release multiple versions of the same track and need precise control over which counts as what, when you operate like a label even if you are technically independent, or when you expect to change distributors and would rather your recording identities not be entangled with a single vendor.

For everyone else, the distributor's prefix is fine — the convenience is real and the cost is zero. Just do not confuse "my distributor assigned it" with "my distributor owns it". Those are different claims, and the second one is false.

When a track needs a new ISRC

The rule that covers every case: if the audio changed, the recording changed, and a different recording needs a different ISRC.

Two of those rows do most of the damage in practice. Assigning a new code to an unchanged master splits one recording's play history into two piles that no chart, royalty statement or analytics dashboard will ever add back together. Reusing an old code for a changed master does the opposite, collapsing two recordings into one identity so that plays of your remaster are indistinguishable from plays of the original.

The instrumental case catches people out because it feels like a derivative rather than a recording. It is not: an instrumental produced by stripping the vocals from your own master is a distinct piece of audio a listener can stream, and it needs its own identifier.

What the code actually does downstream

The ISRC is the join key. Almost every system that owes you money or credit uses it to work out which recording it is dealing with.

Royalty matching. Streaming services report plays against ISRCs. So do collecting societies handling neighbouring rights — PPL in the UK uses ISRCs to track when recordings are played on radio, on television and in public venues, which is income entirely separate from your streaming royalties and from your songwriting royalties. Many independent artists never register for it and never see it. Our guide to making money from AI music covers the income side properly; the point here is just that the ISRC is what makes the matching possible.

Chart eligibility. Charts aggregate streams and sales for a recording across every release containing it, and that aggregation runs on the ISRC. Fragment your identifiers across a single, an album and a re-release, and you fragment the totals.

Content matching. YouTube's rights systems reference recordings by ISRC when matching uploads to reference assets, which is why a clean, consistent identifier matters well before anyone monetises anything. Our YouTube Content ID explainer covers how the matching works and where it goes wrong; the ISRC is the label on the object being matched, not the matching mechanism itself.

Analytics. Platform dashboards deduplicate by ISRC. Two codes on one recording means two rows where there should be one, and every conclusion you draw about performance comes from the wrong denominator. Our distribution guide walks the upload pipeline in which all these fields get filled in.

An identity claim, not a proof of origin

Here is where this page earns its place on this particular site. An ISRC is an assertion about a recording's identity — a number attached at delivery, declared by whoever is doing the delivering. It is the identifier every downstream royalty and detection system uses to talk about your track, and it carries no information whatsoever about how that track was made.

That makes it worth putting alongside the other two things that claim to say "this is this recording", because they work completely differently.

The C2PA standard is the interesting middle column: a signed provenance record that travels with the file and breaks if tampered with, and one of the layers a modern AI export leaves the building carrying. The ISRC is not comparable to it in strength and was never meant to be — it is a filing number, and filing numbers are supposed to be editable by whoever is filing.

Which is exactly why no detection system treats an ISRC as evidence. Distributor classifiers score the audio. They read the third column, not the first.

The layer an ISRC cannot help with

An ISRC gets your track counted. The artifact layer decides whether it gets accepted at all — and if the upload is rejected, the identifier never becomes relevant, because there is no release to identify.

Our standing 50-file corpus of Suno, Udio and Stable Audio output makes the ordering clear. Raw, untreated exports passed 0 of 50 at DistroKid's roughly 0.78 classifier threshold, 3 of 50 at TuneCore's 0.82, and 8 of 50 at CD Baby — where policy rejects fully-AI releases regardless of the score. Perfect metadata moves none of those figures, because the metadata was never what got scanned. Our DistroKid breakdown has the per-platform detail; the methodology page documents how we measure.

The only tool we have benchmarked that addresses the layer those classifiers actually read is Undetectr, which is the first and only AI music watermark remover — the one product built specifically to remove what distributors scan for, rather than a repair suite pointed at the job afterwards. It clears six artifact layers in a single pass: the SynthID-class watermark, the C2PA manifest, the spectral fingerprint and the secondary layers underneath them.

The workflow fits where ISRC assignment already sits in your release schedule. Files go in as MP3, WAV or FLAC, processing runs in the browser with nothing to install, and a track comes back in under a minute — mastered to each destination platform's loudness target in the same pass, Spotify at -14 LUFS and Apple Music at -16. SoundMatch checks for fingerprint collisions before release rather than after a takedown, which matters more than it sounds: two recordings carrying different ISRCs but colliding fingerprints is precisely the shape of a content-matching dispute.

Against that corpus, cleaned files passed 49 of 50 through production distributor classifiers, the single failure being an unrelated copyright flag on a remix-style cover. For comparison, iZotope RX 11 manual restoration scored 32 of 50 at $399 and four to six hours of expert work per track, Audacity workflows managed 8 of 50, and doing nothing scored 0. Undetectr is €39 one-time for the Lifetime tier with unlimited tracks, or €19 for a 10-credit Starter, with no subscription. The full assessment, including where it does not help, is in our Undetectr review.

Honest caveats

Fees and agency processes vary by country and change. The IFPI figure quoted above is an indication, not a quote. Every national agency sets its own terms, and the only reliable source for what registration costs today is the agency covering your territory.

Self-assignment only pays off with discipline. Getting a registrant code is the easy part. Keeping a catalogue log accurate across years of releases, versions and reissues is the commitment, and a badly kept log is worse than none — it produces confident answers that are wrong.

Having an ISRC is not the same as collecting on it. Neighbouring rights, mechanical royalties and performance royalties each require their own registration with their own organisation. The code is the join key those systems use; it is not a claim on any of them.

None of this is leverage. A correct ISRC does not get you playlisted, improve a weak track, or do your marketing. It removes one category of avoidable loss — plays counted against the wrong object, or against nothing at all. Worth getting right, and that is all it is worth.

We are not lawyers or royalty accountants. Everything here reflects standard practice as of mid-2026 and is written for people releasing their own music. If real money or a contested ownership question is involved, you need a music lawyer, not a reference page.