open source audio identification
Identify your music
If you have an unknown music file, one of the applications using AcoustID, such as MusicBrainz Picard or beets, will very likely be able to tell you what song it is, who wrote it, and a lot more.
In case AcoustID doesn't recognize the song yet, you can help by submitting it to the database, so that the next person trying to identify the same song will have more luck.
Get your API key for submitting fingerprints Frequently asked questions
Build it into your application
If you are developing an application that needs to identify unknown audio files, you can use the AcoustID web service. It is completely free for non-commercial applications, all you need to do is register your application.
You can also use the service in commercial applications via AcoustID OÜ, which funds the development of the open source project.
Read the API documentation Register your application Your applications
How it works
At the core of AcoustID is Chromaprint, an open source library for extracting audio fingerprints. It is optimized for matching near-identical audio, which keeps the fingerprints compact and the extraction fast: a two minute recording becomes about 2.5 KB of data, like this one.
The first two minutes of “The Salvation Army” by Dom Um Romão. Time runs left to right, each row of pixels is one bit.
The database stores fingerprints submitted by 1.1M contributors, grouped into AcoustIDs, and links them to 22.2M MusicBrainz recordings. The web service searches it with a custom fingerprint index and answers around 27.5M lookups a day.
All of it is open source, the server included, and the database is available for download.