Music & audio data. Built for your model.
Custom, licensed datasets for AI training. Professional audio, stems and detailed metadata, selected to your requirements.
A broad corpus.
A precise dataset.
- Total music titles
- 13M+
- Total stem assets
- 52M+
- SFX + spoken word
- 5M
- Languages
- 50+
More than 5 million music titles include complete stems and MIDI data.
Beyond the recording / Metadata enrichment
Metadata, in every dimension.
Human-created metadata can be supplemented with automated enrichment across 40+ categories, drawing from thousands of tags and descriptors.
Rhythm & tonality
- BPM
- BPM Range Adjusted
- Time Signature
- Key
Mood & expression
- Mood Tags
- Mood Mean
- Valence Mean
- Arousal Mean
- Energy Level
- Emotional Profile
- Energy Dynamics
- Emotional Dynamics
- Mood Advanced Tags
- Mood Advanced Mean
- Movement Tags
- Movement Mean
- Character Tags
- Character Mean
Instruments & voice
- Instrument Tags
- Instrument Presence
- Predominant Voice Gender
- Voice Presence Profile
- Voice Tags
Genre & era
- Genre Tags
- Genre Mean
- Subgenre Tags
- Subgenre
- Musical Era
- Classical Epoch Tags
- Classical Epoch Mean
Language & discovery
- Description
- Keywords
- Augmented Keywords
- Transformer Caption
- Lyrics transcription
Structure & time
- Mood Segments
- Genre Segments
- Subgenre Segments
- Instrument Segments
- Voice Segments
- Valence/Arousal Segments
- Music Segmentation
- Audio Fingerprinting
Rhythm & tonality4 fields
- BPM
- BPM Range Adjusted
- Time Signature
- Key
Mood & expression14 fields
- Mood Tags
- Mood Mean
- Valence Mean
- Arousal Mean
- Energy Level
- Emotional Profile
- Energy Dynamics
- Emotional Dynamics
- Mood Advanced Tags
- Mood Advanced Mean
- Movement Tags
- Movement Mean
- Character Tags
- Character Mean
Instruments & voice5 fields
- Instrument Tags
- Instrument Presence
- Predominant Voice Gender
- Voice Presence Profile
- Voice Tags
Genre & era7 fields
- Genre Tags
- Genre Mean
- Subgenre Tags
- Subgenre
- Musical Era
- Classical Epoch Tags
- Classical Epoch Mean
Language & discovery5 fields
- Description
- Keywords
- Augmented Keywords
- Transformer Caption
- Lyrics transcription
Structure & time8 fields
- Mood Segments
- Genre Segments
- Subgenre Segments
- Instrument Segments
- Voice Segments
- Valence/Arousal Segments
- Music Segmentation
- Audio Fingerprinting
Automated tagging · Machine-derived metadata
02 / Source & delivery
Structured for your workflow.
Define the file formats, metadata coverage and transfer method your team needs.
Explore formats & metadataSource audio
WAV / FLAC
- Typical sample rate
- 44.1 / 48 kHz
- Typical bit depth
- 16 / 24-bit
Bucket delivery, API or other secure transfer, as agreed.
03 / Rights & licensing
Rights are part of the dataset.
ERISV secures fully cleared master, publishing and other applicable rights for the relevant commercial model-training license.
Understand licensingMaster
Rights in the recording
Publishing
Rights in the composition
Licensed use
Scope for your model
Commercial model training and fine-tuning within the applicable license. Non-commercial R&D licensing can be converted to commercial use.
For catalog owners
Your catalog. Your control.
Access enterprise AI licensing opportunities. Retain ownership, approve uses and restrictions, and participate in revenue under the agreed terms.
For Catalog OwnersParticipation is opt-in. Licensing opportunities are not guaranteed.
Build with ERISV
Start with your requirements.
Tell us what you're training, the audio you need and how you need it delivered.
Contact SalesSource audio
Licensed scope
Licensed dataset