Matnn (Music Audio Tagger Neural Net)

Matnn (Music Audio Tagger Neural Net) is a rest compliant music tag service utilizing a batch queuing system (Kueue) running on a Kubernetes cluster.
Enhance music metadata to improve music discoverability for any use case including music libraries, streaming, song selection.
Matnn utilizes the Discogs-EffNet model from Essentia to predict music classification, BPM (Beats per minute), tonal/key scale and approachability/engagement from over 400 genres of music. Musicnn and Magnatagatune is also offered as a prediction model from 50 genres of music.
To upload your music, use either the simple web form or the CLI driven API service.
Currently music files must be MP3 format and under 10 MB.
Web upload form
Service Offline
Vite App - https://matnn.intamixx.uk/
Direct link without reverse proxy - https://mat.intamixx.uk:8090/upload
Application Architecture
Diagram available here.Submission / Upload API (POST)
Select musical attributes for analysis by specifying tags required. This reduces time taken for analysis to complete. Select either the default 'genre_discogs_effnet' OR 'genre_musicnn' OR 'genre_magnatagatune'. Set just the filename with no options to perform a full set of predictions.
=
=
=
=
A json response will provide the upload status and prediction ID.
There is an average wait of around 30 seconds for a typical prediction to complete. The result will not be available immediately.
Status API (GET)
This will show the status of the prediction using the ID returned from the upload step above.
Result API (GET)
Provides the result of prediction
=
=
Webhook
HTTP-based callback function to desired URL with header signature created from a secret and the payload in attribute X-Matnn-Signature. Body contains original id and event status. Please contact for the webhooksecret key
=
=
Use webhook_url in the request
Webhook received at destination
}
Example Usage
BPM Tag a file
Genre Tag a file
Or use a method / language of your choice.
Contact
Questions? - intamixx@hotmail.com
To Do
- Improve App Frontend
Donate
To help with server costs and administration Buy Me A Coffee ☕