AI Music Resources
Supplements for my classes, workshops, and public demos: generative AI tools, open-source code, and research models for music. I keep adding to and pruning this list so it stays current with what we use in sessions.
- Easyin the browser or a free app, no code
- Moderateneeds an install, a DAW, a MIDI device, or a guided notebook
- Advancedneeds coding, Python environments, or a GPU
- Readinga paper or article, not a tool
My Tools and Code
Open-source AI music tools from my own research, used in workshops and demos.
- ML-Ableton: AI soloist in Ableton LiveAbleton Live sets and a Max for Live device for jamming with a machine-learning soloist. The free Ableton trial works.
- XRJam ML: AI drummer and soloist for virtual avatarsMagenta models that play drums and pentatonic solos for the virtual band in XR Jam, via Ableton and Max for Live.
- AIMAP: the CrowdForge audience-participation systemGenerative multitrack MIDI steered by audience votes. Needs a trained model; contact me for one.
- Icarus: an agent-native DAW for LinuxAn early-prototype open DAW where people, controllers, and AI agents share the same controls. Linux and a Rust build.
Generative AI and Other Tools
Code-Based Open Source
- Google TensorFlow: Music generation tutorial (Jupyter)
- Google Magenta Studio plugins for Ableton Live
- Magenta Colab notebook
- Meta AudioCraft (code, papers)
- Meta AudioCraft (online)
- Meta MusicGen docs (code)
- ACE-Step 1.5 (local song generation)
- YuE (lyrics-to-song)
- HeartMuLa (music foundation models)
- DiffRhythm (full-song diffusion)
- Stable Audio Open (weights)
- Magenta RealTime (live generation)