How to Find Unreleased Tracks DJs Play

An unreleased track is one with no commercial release: a demo, a promo sent ahead of a release date, a white label pressed without artist or title on the sleeve, a dubplate cut for one DJ, or an edit a producer made for their own sets and never put out. DJs mark these in tracklists as ID, which is the honest label for a record nobody has publicly named. If the term is new to you, start with what does ID mean in a DJ set and come back, because the rest of this page assumes it.

The reason this category is large is structural. Labels send tracks to DJs weeks or months before release, and producers road-test their own unfinished material for a year before anyone else hears it. At any given moment a meaningful share of what is being played in clubs cannot be bought by anybody. That is not an obstacle you can search your way past. It is the state of the record.

It is worth being precise about this, because most people burn an hour on it first. A recognizer turns a few seconds of audio into a compact fingerprint and looks that fingerprint up in a database built from released recordings. The lookup succeeds when a matching entry exists.

An unreleased track has no entry. Nobody computed a fingerprint for it, because there was no release to compute one from. The search is not failing, it is returning the correct answer to the question it was asked: this audio is not in the catalogue. Running it through a second recognizer, or a better one, or ours, changes nothing, and any tool claiming otherwise is describing something it cannot do.

This also explains a result that confuses people: a clip where recognition returns nothing at all is often a stronger signal than a clip where it returns something weak. Silence from a fingerprinter on a DJ set frequently means you are listening to something genuinely new. See why Shazam struggles on DJ sets for the other failure modes, including mixed and pitched audio, which are a different problem with a different fix.

This is the method that actually solves unreleased tracks, and it is a reading exercise. Under almost any DJ clip carrying an unfamiliar record there is a thread of people asking for the ID. Somewhere in it is often the answer, sitting next to three wrong ones.

How to weight what you find, strongest first:

1. The DJ replying. The person who played it confirming a name is as close to authoritative as this gets. A heart from the DJ on a comment naming a track is weaker but still meaningful.
2. Several people agreeing independently. Two or three commenters arriving at the same name without quoting each other is strong. Be careful with a chain where everyone is repeating the first reply, which is one guess wearing a crowd's clothes.
3. The producer showing up. On unreleased material the person who made it often appears in the thread, sometimes to say it is theirs and out in March.
4. One confident stranger. A lead worth checking, nothing more. Search the name they gave and listen to it before you believe it. Confidently wrong IDs propagate across threads for years.

The practical problem with this method is not difficulty, it is volume. The answer often arrives days after you looked, buried under a hundred later comments, on a post you have already stopped checking.

Before the comments, read what the DJ wrote themselves. Captions are the most deliberately informative surface on any post, because DJs generally want people to know what they played. Look for an explicit ID: line, a Track: label, an artist - title line, or a bare @producer tag.

That last one is the most underused signal for unreleased material specifically. A DJ playing a track someone sent them will often tag the producer without naming the title, because the title may not exist yet. That tag is still most of the answer: it tells you whose track it is, and from there you follow the producer and wait for the release. On an unreleased record, knowing the artist is the finding.

The full method across post types is in how to find songs on Instagram, and for a single link you want checked now, the Instagram song finder runs the tag, caption and comment checks in one pass.

Here is the thing that changes results more than any technique: unreleased tracks are not solved at the moment you hear them. They are solved later, when the release is announced, when the producer replies to a thread from three weeks ago, or when the DJ posts the same track again with a name attached because it is out now.

That means the winning behaviour is not searching harder today, it is watching the same handful of DJs continuously so you are still there when the answer lands. Doing that by hand is the part that fails. Nobody re-reads a month of comment threads across ten accounts.

This is the specific job CrateWire does: you follow DJs, every new post gets checked, and the posts that do not resolve are not thrown away. As comments accumulate and the DJ or the community adds signals, a post can resolve later without you going back to it. The honest limit is unchanged and worth repeating: if a track is unreleased and nobody has written its name anywhere, no tool here or elsewhere produces one. What changes is that you stop missing the moment when somebody finally does.

For the broader picture of where records come from in the first place, see where DJs find their music.

Once you have a name, or even just a producer, the track is no longer lost, it is pending. Keep it that way deliberately:

• Follow the artist and the label directly on Bandcamp or wherever they sell. Release announcements reach followers first, and small runs sell out.
• Save the post, not just the name. If it turns out the ID was wrong, the clip is the evidence you go back to.
• Do not play a rip. Audio pulled from a video or a mix carries the crowd, the blend and the compression of somebody else's set. It sounds like what it is on a real system.
• Accept the ones that never resolve. Some dubplates are cut for one DJ and never released to anyone. That record staying unbuyable is not a failure of your searching, it is the intent.