Get Started Studio DDSP-VST Demos Blog Research Talks Community

Getting started

Ready to play with Magenta? This page will help you get started making music and art with machine learning, and give you some resources if you want to explore on your own! If you want to watch some conference talks first before you dive in fully, we have a page for that too!


🎸 Musician

If you're a musician, you might want to try out Magenta Studio, our new collection of music plugins, our neural synthesizer (NSynth), or one of our MIDI enabled web applications.

🎨 Artist

If you're an artist, you would enjoy our Sketch RNN demo, or the Quick, Draw! dataset to see what you could build with it.


🕸 JavaScript

Magenta.js is the JavaScript API for doing inference with Magenta models, powered by TensorFlow.js. There are two different libraries currently available:

@magenta/music -- the musical note-based models.

  • try the interactive Hello Magenta getting started tutorial -- this works directly in the browser, without you having to install anything
  • play the demos to see the kind of things you can build
  • browse the API docs
  • see the code on GitHub

@magenta/sketch -- the sketch (drawing) models

@magenta/image -- the image style transfer models

🐍 Python

This library includes utilities for manipulating source data (primarily music and images), using this data to train machine learning models, and finally generating new content from these models.

  • try the Hello Magenta getting started Colab notebook -- if you're not familiar with Colab notebooks, they're a way to set up a Python environment in the browser, so that you don't have to worry about installing it locally
  • use the model-specific Colab notebooks available
  • read the installation instructions
  • see the code on GitHub