BandLab Song Starter is a good way to break writer's block if you are content with composing from a completely random starting point. The composition begins with a roll of virtual dice, the only other parameter is a mood button with morning, day, and night modes. Once you hit a musical idea you like it can be exported to studio mode where it is broken down into stems. There isn't very much depth to this tool but it is incredibly easy to use. BandLab Song Starter it's a beginner-level AI-powered music generator.

BOOMY and AIVA are similar but offer slightly more advanced features. BOOMY is unique in that it actually offers publishing services that allow you to submit music to Youtube, TikTok, and even Spotify with the potential to earn revenue through the platform.

Soundraw (formerly Ecrett Music) is probably the most comprehensive of these kinds of composition solutions in terms of pure composition.The platforms listed above are complete composition solutions for people with limited music ability. Primarily aimed at content creators who want to avoid the music licensing process. Experienced producers or anyone remotely musical may find these tools limiting.

Loudly and Amadeus Topline by Amadeus Code (iOS-based app) are other popular options for random AI-generated music apps.

Separation tools are some of the most exciting AI tools for the music production process. Obtaining stems (the individual recording tracks of vocals and instruments) was until now, exclusively the domain of top-level producers. For the select few who might be lucky enough to work with a record label willing to share the original studio sessions with them.

Vocal separation has been the primary focus but the ability to extract any bassline, guitar lead, piano riff, or drum track from and recording opens the floodgates of potential for sampling. The impact of these tools can't be overstated. Better yet, the ability to download any multitrack from Tracklib, pre-cleared for licensing, and ready for release really is revolutionary. This kind of access to stems was unheard of before now!

LALAL.AI and MOISES.AI are the leading AI separation tools right now. Both offer detailed separation in multiple high-quality audio formats. They differ slightly in their AI engines: MOISES is powered by Deezer's AI and LALAL is powered by their in-house developed neural network, Phoenix. They test differently with different audio inputs and outputs but the comparison is very close sonically.

Atlas by Algonaut and Playbeat by Audiomodern are two groove machine-style VSTs that organize and play your sample libraries in novel ways. Both apps use AI engines to organize your sample libraries into easily customizable and searchable formats within the app. Playbeat even allows you to create your own sample packs from within the interface.

"You can now train the app to learn your preferred patterns with your day-to-day activity, The app then adapts to your own style and it generates personalized patterns to achieve the perfect user experience," according to Playbeat. What really makes these VSTs attractive for a lot of producers are the randomization features.

Atlas' polyrhythm feature can create incredibly complex rhythms. The graphical user interface on these apps is awesome and offers users a new way to visualize and access their sample libraries at once, without ever having to leave your DAW.

These are a new breed of VST and a new way of working for music producers.

LANDR is the leader in online, AI-assisted, mastering solutions. They have a vast reference library at their disposal in order to help it make educated and tasteful decisions when mastering your track, explains LANDR on their site: "Our engine listens to your songs and applies a fast, reliable master without ever using presets." Machine Learning means that AI can get smarter with experience. LANDR has the most experienced AI engine for this.

If you fancy taking your mastering into your own hands there are also great options available for that, too. iZotope is no stranger to AI and machine learning, utilizing the technology in Ozone's Mastering Assistant or the smart features of Neoverb. That said, the RX 10 is their most powerful deployment of neural networks in their tech. The RX 10 uses machine learning to offer smart suggestions on how to clean up and improve audio.

iZotope's audio restoration software works incredibly well, which was also the reason it was featured in our Ultimate Guide to Using Acapella Extractors.

Dreamtonics Synthesiser V is the culmination of years of AI research. The synthesizer replicates the nuances of the human singing voice, allowing users to input lyrics by text and play the melodies they wish the AI to sing them. The deep neural network synthesis engine generates incredibly life-like vocals and is currently unique in its functionality. Dreamtonics Synthesiser V is multi-lingual and offers detailed expressive control over vocals. They have an extensive library of voices, the latest of which, Solaria, is the most advanced and realistic vocal database to date. Synthesiser V is available in AU, VST, and standalone modes.

Orb Producer Suite is a collection of four generative AI plugins designed to create infinite melodies, patterns, and baselines. Orb Chords (a wavetable synth), Orb Melody, Orb Bass, and Orb Arpeggio make up the pack. It features a new polyrhythm engine, block chaining, and lyrical melody features, ideal for AI-generated tracks in a variety of genres.

The Google-backed Magenta project offers a wide range of music AI tools to assist the music production process. Magenta is a machine-learning music project that uses neural networks to help musicians find new ways to express themselves—rather than just simplifying the process. They have invented the world's first hardware neural synth, The N Synth Super. The N Synth Super is built on NSynth (Neural Synthesizer) soft synth technology. It uses a deep neural network to learn the characteristics of sounds, and then create a completely new sound based on these characteristics.

They have also developed Magenta Studio, a collection of music plugins built on Magenta's open-source tools and their model. These tools are available both as standalone applications and as plugins for Ableton Live. If you go to the Magenta home page you can also find a bunch more online tools and you can also participate in demos of their beta products. Best of all, it's completely free!

MuseNet and Jukebox by OpenAI—the developer of ChatGPT—also deserve a mention. "MuseNet is a deep neural network that can generate 4-minute musical compositions with 10 different instruments, and can combine styles from country to Mozart to the Beatles," they write on their website. A variety of genres can be combined in novel ways in MuseNet. Yet classical music and its relationship with popular music seem to be the foundation.

Jukebox is a neural net that generates music, including rudimentary singing, as raw audio in a variety of genres and artistic styles. It should be noted that these are code-based programs, not apps—for now. So this is more the domain of coders. There are detailed instructions if you want to try it out and the fact it's all open-source is very cool.