BRAVE Plugin

Reimagining Live Instrumental Performance.

Get the Plugin Training Code

Working with sound or interaction? Let's make better tools for performers. Share your Insights

Rediscover your Voice and Instrument

Instant Live Transformation

Experience super-low latency timbre transfer for seamless, on-the-fly timbre transfer as you play.

Natural and Intuitive

Interact effortlessly with advanced AI models, making complex sound manipulation feel malleable and accomodating.

Capture what's Important

We capture subtle nuances and articulations, translating them into dynamic and expressive sonic output.

No matter what you play

Try enhancing your instrument's capabilities with new sonic dimensions while keeping your unique playing style.

See BRAVE in Action

Play live, in time

Percussive augmentation

Beatboxing

Instrument retargeting

Get the Plugin & Pretrained Models

Download the latest version for macOS Apple Silicon and the essential pretrained models to get started.

Download Plugin Download Pretrained Models

Plugin notarized by Apple - Latest update: `6/3/25 - Version 1.0`

Note: This plugin is currently in beta testing, and there may be bugs or unexpected behavior. Your feedback is appreciated to help improve future versions. If you encounter any issues, please contact us.

Instructions

Use your preferred DAW (Digital Audio Workstation), that supports VST3 or AU plugin formats. Each pretrained model implements a new sound you can control with your instrument.

  1. Download the plugin to a directory where your DAW can see it.
  2. Download and uncompress the pretrained models.
  3. Open the plugin at your DAW, and select the folder where you stored the models.
  4. Done! The models can now be selected from the GUI's list.

Pretrained models

The provided models are trained from openly available datasets.

Model Dataset Description
birdsong.h5 British Birdsong Dataset, gathered from the Xeno Canto collection.
candombe.h5 Uruguayan candombe drumming dataset.
drumset.h5 The Expanded Groove MIDI Dataset. 2h 50m of a drummer performing on an electronic drum kit.
filosax.h5 Trained using recordings of Participant 1 from the Filosax Dataset.

Check BRAVE's Repo for instructions on how to train your own models. Contact Me if you would like to submit new models to the pretrained pack.

Contact

If you want to know more, stay up to date on this project, report a bug, or have any feedback, please contact us.

Contact Form