What we do
Musicovery is a high quality and comprehensive music recommendation engine, very easy to integrate through its API.
It provides 4 types of services:
- descriptive metadata on artists and tracks (genres, moods, era, geographic, acoustics descriptors…)
- recommendations and playlists, personalized in real time
- bespoke webservices to provide specific content (recommendation of live concerts, recommendation of playlists, Youtube channels,…)
- advice on data analysis, algorithms, recommendation optimization, metadata sourcing and music UX design
With more than 12 years of experiments on how to provide intuitive, rich and smart radios, how to make sense of behavioural data and produce precise descriptive metadata on music content, Musicovery is in a unique position to provide a comprehensive recommendation engine. It generates any kind of recommendations and playlists: from a mood, a song, an artist, a subgenre, a theme, a place, or for a specific listener.
Musicovery measures the quality of recommendations and playlists with an analytic tool that optimizes recommendations and playlists to each listener.
Recommendations and playlist are provided through an API, very easy to integrate, especially for prototyping new innovative UX.
Musicovery provides metadata on:
- tracks :
- moods: key words like "happy", and mood quantitative value (valence/arousal)
- activities : listening situations like music for driving, working, partying…
- acoustic descriptors
- artists : genres, era, role, geographic location
Metadata are attributed by experts (for the top of the catalog and archetypical songs of each genres/style/moods) and automatically by machine learning from audio and semantic web (for the long tail of catalogs).
Musicovery indexes systematically the commercial catalog and beyond you can upload your audio files to Musicovery API to get the corresponding metadata.
Moods and activities indexation are the result of extensive researchs conducted on psychological states mapping (circumplex models like Russell, Plutchik,...), their mapping with acoustic descriptors, and on automatic indexation by machine learning (R&D projects with Ircam on music information retrieval).
- tracks :
Musicovery API generates the best playlists and recommendations from a mood (calm, happy,...), an artist, a track, a genre/style, a context/activity (for driving, working, partying,...), a theme, a period/year, a location (city, region, country, continent).
Recommendations of tracks, artists, genres and playlists are personalized in real time to each listener, according to his music preferences, listening behaviour and listening history.
With very few information on a listener music preferences and listening behavior, Musicovery engine starts very early to personalize recommendations and playlists with a high degree of relevance.
Try a playlist:
- Launch radio from seed artist Coldplay
- Try a radio Classical symphonic
- Try radio year 1969
- Try radio calm
- Try radio San Francisco
Playlists and recommendations can be restricted to a specific catalog, and be optimized for a specific UX and a specific audience
Musicovery API makes it very easy to provide descriptive metadata, recommendations and playlists in real time.
You can test the API freely and look at the results returned by the API for the following example:
Ex. : get a playlist from Skrillex, with artists little known from the same genre (dubstep)
To learn about all the functionalities provided by Musicovery, please read the documentation API.
WEBSERVICES & ADVICE
Musicovery provides also services like personalized and geolocalised recommendation of live concerts, recommendation of playlists, personalized Youtube channels, emerging artists for specific regions and genres…
These services require to identify relevant sources of data and content partners and to map identifiers of the content partners.
Musicovery sets up bespoke webservices tailored to the specific needs of its clients.
To make the most of music recommendation services, Musicovery provides its clients with advice on:
- Recommendation optimization (data analysis, algorithm design, recommendation quality measurement, A/Z tests, algorithm benchmark)
- Music content and metadata sourcing
- Music UX design and dataviz
BlogAfter investigating music recommendation from the perspective of music platforms product, listeners behaviour, recommendation systems design, we are going to consider where recommendation fits within music platforms overall strategy. Beyond an initial strategic positioning, the ultimate shape...April 18, 2019In this article we are going to look at how algorithms can help personalize recommendations to various types of listeners. Although a recommendation system is a tryptic device/UI/algo, we will focus here on the issues raised by algorithms and consider the following features: user-to-items ...March 29, 2019The ultimate goal of listening to music is pleasure. But first listeners need to have a reason to discover a new song, want to listen to it again. After several repetitions they start liking the song, and getting pleasure. Once they have listened to it too often, they reach a saturation point,...
Keep in touch
To get the last news about Musicovery technology, subscribe to our newsletter.