Construction Kit Graph
Every demo mix is connected to its loops, one-shots, BPM, key, and instrument categories.
MIX
DRUMS
BASS
MELODY
TOP LOOPS
DEMO MIX
PERCUSSION
BPM
KEY
FX
VOCALS
Audio datasets.
Commercial AI Training.
Train music AI with legally licensable proprietary audio.
80,000+ original sounds structured for training, fine-tuning, source separation, and music generation.
100%Proprietary catalog
Daily File UpdatesSounds are uploaded daily.
METADATAEmbedded directly in each audio filename
DATASET ONLINE
SML-X
Construction kits
0
Original sounds
80K+
Daily File Updates. Sounds are uploaded daily.
Multi-genre catalog Dataset for AI Training
01
Dataset intelligence
A structured music system.
Training-Ready Metadata
As shown below, unique ID, instrument, BPM and key are written directly into each audio file name, keeping the metadata attached to the file itself.
METADATA
"kit_id": SML_TRP1
"instrument": Bass_Loop
"bpm": 126
"key": Fm
Example WAV File:
SML_TRP1_Bass_Loop_126bpm_Fm.wav
100% Owned Audio Catalog
Every sound is created in-house and fully owned by Smokey Loops, original, high-quality audio with a clear chain of title and no third-party sample restrictions.
100%Proprietary
24-bitHigh quality
0Third-party claims
✓Clean rights · AI-ready · Commercially licensable
Active Genres
A continuously expanding multi-genre catalog to support specialized and general-purpose music AI models.
- Balanced coverage across electronic, urban and cinematic categories
- Genre-specific kits, loops, one-shots, vocals and sound-design assets
- Custom genre mixes available for targeted training objectives
Daily Updates
New proprietary files are added every day, keeping licensed datasets current and expandable throughout the active term.
- Daily ingestion of newly produced audio assets
- New construction kits, loops, one-shots, vocals and metadata
- Quality control, filename validation and catalog organization
- Eligible updates included according to the active license plan
Catalog expansion in progress
Flexible Delivery
Dataset delivery is adapted to the client’s infrastructure, security requirements and technical ingestion workflow.
- Secure cloud delivery or private download area
- Dropbox, Google Drive or approved enterprise storage
- Encrypted hard drive shipment for large datasets
- Upload to private FTP, SFTP, server or client-managed bucket
- Custom folder structure, manifests and checksum files available
Dataset architecture
From the demo mix
to every single element.
The structure preserves the original musical context. Every asset can be used individually or as part of a relational system built for AI pipelines.
- Complete demo mixes
- Synchronized loops
- Categorized loops
- Metadata embedded directly in every audio file name
- Licensing
- 100% Proprietary catalog
▾ SMOKEY_AI_DATASET Example
▾ TECH_HOUSE_PACK_018 Example
▾ CONSTRUCTION_KIT_01
♪ SML_TH1_Drum_Loop_125bpm_F
♪ SML_TH1_Bass_Loop_125bpm_F
♪ SML_TH1_Synth_Loop_125bpm_F
♪ SML_TH1_Vocal_Loop_125bpm_F
♪ SML_TH1_Perc_Loop_125bpm
♪ SML_TH1_Kick_Loop_125bpm_F
♪ SML_TH1_Clap_Loop_125bpm
♪ SML_TH1_Hihat_Loop_125bpm
♪ SML_TH1_Fx_Loop_125bpm
♪ SML_TH1_Demo_Mix
02
Pack audio previews.
Audio catalog
Pack audio previews.
AUDIO PREVIEW
Mobile previews are currently unavailable.
For the best listening experience, please open this page using a desktop or laptop web browser.
Page 1
00:00
00:00
◖
03
Choose the dataset scope
AI Licensing Programs
Choose the dataset scope
and license the AI use.
Start with non-commercial research or select an active commercial AI license. Commercial plans include 12 months of authorized use, dataset updates and support. Renewal keeps commercial AI rights active and is quoted separately.
01RESEARCH
RESEARCH & R&D
100 Kits
For very small startups, academic teams and independent labs developing prototypes, proofs of concept and internal research with no commercial use.
original sounds≈ 2,000–3,000+
from1,490 €
- 100 proprietary construction kits with demo mixes, loops and one-shots (or individual sounds)
- Filename-level metadata including kit ID, instrument, BPM and key
- 12-month AI research license for internal R&D, testing and proof-of-concept work
- One internal research model or prototype
02STARTER
DEVELOPER
250 Kits
A focused proprietary dataset for first-generation commercial music AI products and approved production prototypes.
original sounds≈ 6,000–8,000+
from7,900 €
- 250 complete construction kits with demo mixes, loops and one-shots (or individual sounds)
- Filename-level metadata including kit ID, instrument, BPM and key
- 12-month active commercial AI license for 1 approved model
- Commercial deployment rights for the agreed product
- Annual renewal required to continue commercial AI operations, training and deployment (quoted separately)
03PRO
PROFESSIONAL
500 Kits
For production teams running multiple experiments, fine-tuning pipelines and commercial releases across several music categories.
original sounds≈ 12,000–15,000+
from13,900 €
- 500 complete construction kits selected across multiple genres (or individual sounds)
- Filename-level metadata including kit ID, instrument, BPM and key
- 12-month active commercial AI license for up to 3 approved models
- Commercial deployment rights for the agreed products
- Updates, onboarding and eligible catalog releases included during the active term
- Annual renewal required to continue commercial AI operations, training and deployment (quoted separately)
04SCALE
BUSINESS
1,000 Kits
A broad proprietary training base for growing platforms, internal model portfolios and production-scale music generation workflows.
original sounds≈ 24,000–30,000+
from23,900 €
- 1,000 complete construction kits with relational musical context (or individual sounds)
- Filename-level metadata including kit ID, instrument, BPM and key
- Expanded multi-genre selection for broader model coverage
- 12-month active commercial AI license across multiple approved internal models
- Commercial deployment rights for approved applications
- Priority support, updates and eligible catalog expansion included during the active term
- Annual renewal required to continue commercial AI operations, training and deployment (quoted separately)
05FULL CATALOG
COMPLETE LIBRARY
≈ FULL LIBRARY
Maximum proprietary catalog coverage for enterprise-scale training, foundational research and large commercial AI deployments.
original sounds≈ 80,000+
from69,900 €
- Complete proprietary audio library
- Filename-level metadata including kit ID, instrument, BPM and key
- Approximately 4,000 construction kits and 800+ original packs
- Approximately 80,000+ original sounds across demo mixes, loops and one-shots
- 12-month active enterprise AI license for the agreed model and product portfolio
- Enterprise support, updates and eligible new releases included during the active term
- Annual renewal required to continue commercial AI operations, training and deployment (quoted separately)
- Custom delivery structure based on technical ingestion requirements
The Research plan is strictly non-commercial. Any monetization, customer-facing release, commercial deployment or paid service requires an upgrade to an active commercial AI license. Commercial prices are indicative starting prices and exclude VAT where applicable. Renewal pricing is quoted separately according to model scope, deployment scale, support requirements and dataset expansion. Final rights and restrictions are defined in the signed agreement.
License model
Commercial non-exclusive
AI licensing
05
Why companies choose
Built for serious AI teams
Why companies choose
Smokey AI Dataset.
A proprietary music catalog combined with structured organization, practical delivery options and licensing designed for real AI development workflows. For over 10 years, creating exclusive sounds for the music industry.
Proprietary audio only
Original audio created and controlled by Smokey Loops, with a clear chain of title and no dependency on third-party sample libraries.
Structured metadata
Kit ID, instrument, BPM and key remain attached to each asset through consistent filenames and organized folder structures.
Relational music context
Demo mixes, loops, stems and one-shots preserve their musical relationships instead of being delivered as an unstructured file dump.
Daily catalog expansion
New proprietary sounds are produced and added continuously, allowing licensed datasets to evolve during the active term.
Enterprise delivery
Receive data through secure cloud storage, encrypted drives, private FTP or SFTP, client servers or managed storage buckets.
Commercial AI licensing
Plans define dataset scope, authorized model usage, deployment rights, updates and support for commercial AI operations.
Frequently asked questions
Technical, licensing
and delivery answers.
Key information for research teams, AI startups, product companies and enterprise procurement.
Ask a specific questionCan the dataset be used to train or fine-tune AI models?
Yes. The permitted training, fine-tuning and deployment scope depends on the selected plan and the signed license. Research use is non-commercial; commercial plans authorize approved commercial AI operations during the active license term.
Do clients purchase ownership of the audio files?
No. Smokey Loops retains ownership of the audio and underlying intellectual property. Clients receive defined usage rights for the dataset and authorized AI models under the applicable agreement.
Can the original dataset be redistributed or resold?
No. Raw audio files, stems, loops, one-shots, metadata and dataset structures may not be redistributed, sublicensed or resold unless a signed enterprise agreement explicitly permits it.
What metadata is included?
Core metadata can include kit ID, instrument or asset category, BPM and musical key directly in filenames. Folder maps, manifests, checksums and additional exports can be included according to the project scope.
Which audio formats and quality are available?
The catalog is prepared as high-quality production audio, commonly including 24-bit WAV assets. Final format, sample rate, channel configuration and delivery specifications are confirmed for each licensed dataset.
How are large datasets delivered?
Delivery can use secure cloud storage, Dropbox, Google Drive, encrypted hard drives, private FTP or SFTP, client-managed servers or storage buckets. Folder structure and verification files can be tailored to the ingestion workflow.
Are new catalog updates included?
Eligible updates are included according to the active plan and agreement. The initial commercial term is 12 months; continued access, additional training and ongoing commercial operations require renewal where specified by the license.
Can a plan be upgraded or customized?
Yes. Dataset size, genre balance, instrument categories, metadata, delivery method, model scope and support can be expanded through an upgrade or tailored enterprise proposal.
Can you build a dataset around specific genres or instruments?
Yes. Custom selections can focus on genres, instruments, vocals, sound effects, loops, one-shots or construction kits, subject to catalog availability and the agreed technical requirements.
How does annual commercial renewal work?
Commercial plans provide an active 12-month AI license. Renewal terms are quoted according to model scope, deployment scale, support and dataset expansion. The signed agreement governs continued commercial use and post-expiry obligations.
Private dataset access
Request a private demo, a sample dataset, or a tailored enterprise proposal.
license@smokeydatasetai.com
info@smokeyloops.com