AcousticBrainz

dataset·crowdsourced music features·active

Crowdsourced low- and high-level acoustic descriptors indexed by MusicBrainz recording identifiers.

Recorded facts

Official site	https://acousticbrainz.org/ ↗
Geography	Spain
size	submissions: 29460584 · unique recordings: 7564215
scope	Large-scale reusable music audio features without redistributing source recordings.
genres	multiple
license	CC0
steward	Music Technology Group and MetaBrainz
languages	multilingual metadata
modalities	JSON low-level features; JSON high-level predictions; CSV feature dumps; MusicBrainz IDs
geographies	Global
access method	Website API and downloadable dumps.
intended uses	music classification; similarity; feature research
annotation method	Essentia signal analysis and pretrained high-level classifiers.
collection method	Community clients computed Essentia features locally and submitted outputs.
provenance claims	Features are indexed by MusicBrainz recording IDs; original audio is not redistributed.
current availability	Collection stopped in 2022; website, API and dumps remain available.
commercial use limits	No dataset-level restriction under CC0.
us market scope basis	Publicly available or materially used in US-facing music/audio-AI research.
limitations biases disputes	High-level features are less stable than low-level descriptors.; Crowdsourced submissions include duplicates and uneven catalog coverage.
redistribution restrictions	None claimed for AcousticBrainz feature data.
related models tools benchmarks	Essentia; MusicBrainz; AcousticBrainz Genre Dataset
corrections withdrawals replacements access changes	Stopped accepting submissions in 2022; downloadable archives preserved.