ISRC codes explained: what they are and how to use them in your music app

Every commercially released recording carries a unique code that identifies it across every platform, database, and royalty system worldwide. It isn’t the audio fingerprint used for recognition — it’s a 12-character alphanumeric identifier called the ISRC.

If you’re building music apps, managing catalogs, or tracking royalties, you’ll deal with ISRCs constantly. This guide explains what they are, how they’re structured, and how to retrieve and validate them programmatically.

What is an ISRC code?

An ISRC (International Standard Recording Code) is a unique identifier assigned to a specific sound recording. The displayed form looks like US-ABC-12-34567 and breaks down into four parts:

US — country code (2 characters)
ABC — registrant code (3 characters, assigned to the label or distributor)
12 — year of reference (2 digits)
34567 — designation code (5 digits, unique within that registrant and year)

Each ISRC identifies one specific recording. The same composition recorded by different artists gets different ISRCs, and even a re-recording by the same artist gets a new ISRC. The hyphens are for display only — stored and transmitted, the code is 12 characters with no separators (USABC1234567).

Why ISRC codes matter for music apps

Titles and artist names are not dependable keys. “Hurt” could mean Nine Inch Nails’ original or Johnny Cash’s cover. “The Beatles” might be stored as “Beatles” in one database and “The Beatles” in another. An ISRC removes that ambiguity — it points at exactly one recording regardless of how the surrounding metadata is formatted.

Common use cases

Royalty tracking — performance rights organizations and platforms report plays by ISRC, so accurate codes are tied directly to whether rights holders get paid.
Catalog management — labels use ISRCs as the backbone for tracking recordings across territories, formats, and platforms.
Rights and copyright — content-identification systems link audio to ownership information through the ISRC.
Analytics — one ISRC equals one recording, which is what makes cross-platform play aggregation possible.

How ISRC codes are assigned

The International Federation of the Phonographic Industry (IFPI) administers the ISRC system globally. National agencies assign a registrant code to each label or distributor, who then generates unique ISRCs for their recordings. In practice the flow is: register with a national agency, receive a registrant code, generate ISRCs per recording, embed them in files and database entries, and report plays using them.

Getting ISRC data through an API

Music recognition APIs can return the ISRC alongside other metadata once they identify a recording from audio. With AudD, the ISRC is available on the Startup plan and higher — on lower plans the field is simply absent from the response, which is normal and not an error.

from audd import AudD

audd = AudD("your-api-token")  # get a token at dashboard.audd.io

song = audd.recognize("https://audd.tech/example.mp3")
if song:
    print("Title:", song.title)
    print("Artist:", song.artist)
    print("ISRC:", song.isrc)  # None on plans below Startup, or if unknown

Treat the ISRC as nullable. A successful match can lack one — older recordings and some independent releases were never registered — so design your code to handle its absence rather than assuming it’s always present.

If the SDK doesn’t model a particular field as a typed property, you can still read it: in Python any extra field lands on the result’s model_extra map, and in Node on extras.

import { AudD } from "@audd/sdk";

const audd = new AudD(process.env.AUDD_API_TOKEN ?? "test");

const song = await audd.recognize("https://audd.tech/example.mp3");
if (song) {
  console.log({ title: song.title, artist: song.artist, isrc: song.isrc });
}

Validating and normalizing ISRCs

A valid ISRC is exactly 12 characters: a 2-letter country code, a 3-character alphanumeric registrant code, a 2-digit year, and a 5-digit designation code, with no separators in storage. A few rules keep your data clean:

Normalize on input. ISRCs arrive with hyphens (US-ABC-12-34567) or without (USABC1234567). Strip separators and uppercase before storing, then re-insert hyphens only for display.
Validate the format. Reject codes that don’t match the 12-character pattern so malformed data doesn’t enter your catalog.
Plan for missing codes. Not every recording has an ISRC. Keep a fallback identification path (audio recognition, or artist/title matching) for records that lack one.
Don’t assume global uniqueness is perfect. Human error occasionally produces a duplicate ISRC. If an ISRC is your primary key, log collisions rather than silently overwriting.

ISRC and audio recognition together

Audio recognition and ISRC retrieval complement each other. Recognition answers “what recording is this?” from the audio; the ISRC it returns gives you a standardized identifier you can hand to royalty, rights, and analytics systems — without running a separate lookup. That combination, audio-based identification plus a stable cross-platform code, is what lets an app go from a raw clip to a record that the rest of the music industry already understands.

If you need to identify every recording in a long file rather than a single clip, the enterprise endpoint chunks the audio server-side and returns every match with timestamps (and ISRC on Startup-and-above), billed per 12 seconds.

matches = audd.recognize_enterprise("long-mix.mp3", limit=10)
for m in matches:
    print(m.timecode, m.isrc, m.artist, "—", m.title)

Set a limit during development so a long file doesn’t expand into more billed work than you intend.

Wrapping up

ISRC codes are the identifiers the rest of the music industry already speaks. Retrieve them from recognition responses, normalize and validate them on the way in, and design for the cases where they’re missing. A recognition API that returns the ISRC alongside the audio match gives you both halves — identification and a standardized key — in a single call.

Get a token at dashboard.audd.io and read the API reference to start.

Related