Hear the signal behind the song.

What Counts As AI Music?

AI music is any recording whose audio was produced, in whole or in part, by a generative model rather than by performers, instruments and microphones. In 2023 that mostly meant novelty loops. Today a single text prompt can return a finished three-minute song with lyrics, a lead vocal, backing harmonies, a full arrangement and a competitive master. Millions of such tracks are uploaded to streaming services every month, which is why so many listeners now find themselves asking a question that would have been absurd a few years ago: is this song AI generated?

It helps to separate three very different things that all get called "AI music". First, fully generated tracks, where a model produced the entire recording end to end. Second, AI-assisted production, where a human wrote and performed the music but used machine tools for stem separation, mastering, pitch correction or arrangement ideas — this now describes a large share of all commercial releases and is entirely unremarkable. Third, synthetic vocals over human instrumentation, or the reverse, where the two are blended within a single song.

Only the first category is what most people mean when they ask whether a track is AI generated, and it's the only one this tool tries to identify. The distinction matters, because AI-assisted music isn't dishonest and isn't unlawful. A tool that treated every use of a machine tool as a red flag would be useless — it would flag most of the charts.

What happens after you choose a file

The public detector follows a short server-side path. Your browser checks the file, reads its duration and creates a content hash. The complete audio then travels over HTTPS to our server, which validates the request and looks for a recent result for the identical file.

1. 1. Request check

   The server enforces the 25 MB and 5-second-to-15-minute limits, then applies the 12-analysis fair-use allowance.

2. 2. Cache check

   A numeric result for the same file, provider, model and integration version may be reused for up to 14 days.

3. 3. Specialist classification

   When no valid cache entry exists, the audio is sent to a third-party music-classification provider. This site does not run a browser model to produce the verdict.

4. 4. Report assembly

   The provider's categorical verdict, component indicators, confidence bands and timeline are normalised for the report. The application does not calculate a blended AI percentage.

Uploaded audio remains in application memory only for the request and is not written to our database or file storage. The provider handles its copy under its own terms. Read the documented request lifecycle for the complete boundary between the browser, this service and the provider.

AI vs Human Music

There is no single property that separates generated music from recorded music, which is precisely why detection is hard. What exists instead is a set of tendencies, each with enormous overlap between the two populations. Understanding those tendencies — and their exceptions — is the difference between using a detector well and misusing it.

Now the exceptions, because they're the whole story. Loudness-maximised electronic music, template-driven pop, quantised programming and low-bitrate uploads all produce exactly the "generated" signature above, despite being entirely human work. Meanwhile a generated track that's been re-recorded through speakers, re-mixed with live overdubs, or simply produced by a newer model with wider bandwidth will look comfortably human. False positives and false negatives aren't edge cases here — they are the normal operating condition of every acoustic detector currently available, including this one.

Generator context—not attribution

These are the platforms whose output shaped the measurements this tool takes. Read this list carefully: it means the tool was designed with this kind of audio in mind, not that it can identify which platform made your file. The current model carries no per-generator labels, so it never reports "made with Suno". When the evidence leans generated it says Unknown AI generator, and when it doesn't it says Likely human recording.

• Suno

  A prompt-to-song service that can return vocals, lyrics, arrangement and a finished mix.

  The detector accepts its exports but cannot verify or attribute them to Suno.

• Udio

  A music generator with extension and inpainting workflows for building longer arrangements.

  Section changes may be useful context to a listener, but they do not identify Udio.

• ElevenLabs Music

  A music-generation product from a company also known for synthetic voice tools.

  Provider-returned vocal and instrumental indicators are not generator labels.

• Stable Audio

  A generation service oriented toward instrumental music, loops and sound design.

  Bandwidth or codec clues are ambiguous and cannot establish platform origin.

• Riffusion

  A generative music system whose earlier work used spectrogram-image diffusion.

  Historical artefacts are not a dependable fingerprint of current output.

• Mubert

  A generative service for background and production music.

  Looping and repetition also occur throughout human-made production.

• Seed Music

  A research-lineage system for generating vocals and accompaniment together.

  This site has not measured a platform-specific detection rate.

• MiniMax

  A multimodal model family that includes music-generation workflows.

  Reference-conditioned output cannot be identified from the final waveform here.

• Mureka (Sonauto)

  A prompt-to-track service with vocal and style controls.

  A stylistic resemblance is not a platform attribution.

• Other and future models

  Generators not listed here, including products released after this guide was written.

  The public result does not depend on selecting a generator name.

Naming the platform behind a track is a genuinely harder problem than detection, and it degrades even faster as models are updated. It's on the roadmap only behind a properly evaluated classifier. Until then, any tool that confidently names a generator from audio alone is telling you more than it can actually know.

Why Check A Track At All

People reach for a tool like this for very different reasons, and the appropriate strength of evidence differs enormously between them. Casual curiosity needs almost nothing; a takedown needs far more than any acoustic tool can offer.

Listeners

Checking an unfamiliar track before sharing it, adding it to a playlist or recommending it to someone else.

Musicians

Reviewing a suspicious collaboration, sample pack or submission before signing anything or releasing it.

Labels and curators

A first, non-binding screening step in front of a human review — never the review itself.

Journalists

Adding one technical data point to a story that already rests on human sourcing and documents.

Educators

Demonstrating concretely how synthetic-media detection works, and — more usefully — where it fails.

Researchers

Inspecting reproducible, documented signal measurements on their own material rather than a black-box score.

Worth stating plainly: checking a track isn't about punishing anyone. Generated music is legal, and much of it is made by people who are open about how they made it. The legitimate uses here are transparency, honest labelling, platform moderation policy and research — not vigilante accusations built on a single percentage.

Limitations

This section is longer than most competitors' because it's the most useful part of the page. Here's what this tool genuinely cannot do.

• It cannot name the generator behind a track, and it will never guess one.
• It does not separate stems. Vocal and instrumental indicators are labels returned by the provider, not independent analyses performed here.
• It cannot prove authorship, and it is not forensic evidence in any legal, academic or employment context.
• It cannot see through heavy remixing, re-recording, re-amping or aggressive mastering.
• It cannot recognise a specific song, artist or release by fingerprint — it has no database.
• It cannot reliably assess very short clips; under ten seconds there is simply not enough material to sample.
• It cannot compensate fully for low-bitrate encoding, which removes the exact detail the analysis reads.
• It cannot keep pace automatically with generators released after the current model version.

Detection is also inherently asymmetric. A tool like this is much better at raising support for investigation than as a clearance certificate. A not-AI verdict records what the provider found in this file; it does not reconstruct the production process. Treat it as one observation, especially when the audio has been heavily edited. The limitations page lists every failure mode we know about, including why we don't publish an accuracy percentage.

The data boundary

The complete audio travels over HTTPS to our server and the specialist provider. Our application keeps it in memory during the request and does not write it to site storage. The provider handles its copy under its own terms.

Numeric output may be cached for up to 14 days against a file hash. The cache is not an account history or public report. A salted one-way network-address hash enforces fair use; the raw address is not stored by that mechanism.

No shareable result URL is created. Generate the PDF in your browser if you need a record; it excludes provider and request identifiers. Read the privacy policy.

Explore the detector

• Is This Song AI Generated?
• Accuracy and limitations
• Benchmark protocol
• How to Tell If a Voice Is AI Generated
• Blog — guides and research notes
• How AI Music Detection Works