SoundCloud has started rolling out a new way for people to discover music and audio files on its service. Called Suggested Tracks, this feature analyzes your listening activity to surface the tracks and artists you may not have heard before. As the war among music services continues, this is SoundCloud’s answer to Spotify’s Discover Weekly offering.

With Suggested Tracks, SoundCloud said it uses machine learning to pass along “fresh new” music and audio suggestions from among its 125 million tracks in its library. Of course, the more you use the app, the better it can adapt to your listening preferences. This capability is available on both the web and on SoundCloud’s iOS and Android apps.

Existing users will already have suggestions waiting for them — these listings will be available under the Discover tab when using SoundCloud on the web or in the search bar when on mobile. The company said that new users will start seeing suggestions once more tracks have been listened to or liked.

The release of this feature comes more than a week after SoundCloud received a $70 million investment from Twitter. In the past year, it seems to be beefing up its offering to provide a better experience to users, especially because there’s no shortage of offerings, such as Apple Music, Pandora, Slacker, and Spotify, which recently announced it has 100 million users, 30 million of which are paying subscribers.

In March, SoundCloud began its new $9.99 monthly subscription music service in the hopes of generating more revenue. The idea was to offer a premium service that lets you listen to any song in the company’s catalog that you want, similar to what you’d get with Spotify’s premium service. It also made its service international-friendly, starting with customers in the U.K. and Ireland this year.

One thing that may differentiate SoundCloud from the others is the push toward independent artists and creators. The company launched a mobile app for these users last year that enables them to better manage their accounts and community. With the roll-out of a learning discovery feature, this could improve the odds of their work being found and listened to instead of music from someone with a larger brand who has more resources at their disposal.