Back to browse
GitHub Repository

The Infinite Crate is a DAW plugin built on JUCE, React, and the Lyria RealTime live music model

93 starsTypeScript

An open-source DAW plugin built on JUCE, React, and Lyria RealTime

by DesaiAshu·Mar 9, 2026·2 points·0 comments

AI Analysis

●●●BangerWizardryZero to One

Steer live AI music generation with text prompts inside your DAW.

Strengths
  • Lyria RealTime enables genuine real-time audio streaming not batch generation
  • Text prompt mixing provides granular control over AI output
  • Production-ready with VST3, AU, and Standalone builds already shipped
Weaknesses
  • Requires Google's Lyria API dependency limiting offline use cases
  • Niche market of producers comfortable integrating AI tools into workflows
Category
Target Audience

Music producers, audio engineers, live performers

Similar To

Google Magenta · AIVA · Soundful

Post Description

Hi HN!

I'm a former YC founder/visiting partner and independent musician/producer in SF. About a year ago I started collaborating with the Magenta research team as an artist to explore the role of realtime generative music models in studio production and live performances

Through our early experiments we recognized the need to bridge the gap between server based generative music models and professional audio software, and to design interfaces for generative models that are more steerable and controllable than the current "push button to generate song" paradigm. We built and released a DAW plugin which has used by some of our favorite artists — including a wonderful showcase with Daito Manabe in Tokyo — and was featured as an exciting new music tool at NAMM

Today we're open-sourcing the plugin under Apache 2.0 for the community to fork and build new interfaces with

The plugin is architected using React/TypeScript for the UI layer and JUCE/C++ for DAW connection, audio processing, and websocket audio streaming from the Gemini/Lyria API. This allows rapid iteration on the frontend using hot-reload (Shadcn/Tailwind), while ensuring latency sensitive operations (audio streaming and playback) happen in a compiled and unmanaged language with a tight clock. State is synced between TypeScript and C++ using Zustand’s state management and nlohmann json

The JUCE + React structure was inspired by Spencer Salazar's talk at ADC in 2020 (https://www.youtube.com/watch?v=XvsCaQd2VFE) and hopefully can help make plugin development easier web developers and vibe-coders

We'd love to get your input and ideas for how to improve on it. In the near term, we hope to update the plugin to support on-device inference of the Magenta RealTime open-weights music model for offline use. In the long term, we hope to support future music models with improved controls, such as audio and MIDI input

Cheers!

Similar Projects

AI/ML●●●Banger

Magenta Real-Time Music Generation on iPhone, Without the GPU

Splits a 230M-param model across ANE and CPU to avoid thermal throttling on iPhone.

WizardryBig BrainZero to One
MediaSquirrel
904d ago