DAWZY: Human-in-the-Loop Natural-Language Control of REAPER

Aaron C Elkins, Sanchit Singh, Sawyer Blankenship, Adrian Kieback, Uyiosa Philip Amadasun, Aman Chadha

Keywords: Digital Audio Workstation (DAW), Natural Language Interface, Music Production, Human-AI Collaboration, Code Generation, Interactive Music Systems, Music Technology, Creative Workflow, Large Language Models (LLM), Model Context Protocol

TL;DR: Music Production via AI interface

Abstract:

We demo DAWZY, a voice- and text-driven assistant that translates natural-language requests into
reversible, state-aware actions in REAPER, a scriptable Digital Audio Workstation (DAW). DAWZY
shortens the gap between musical intent (e.g., ‘warm the vocals’) and precise DAW edits, keeping the
DAW as the creative hub. The system queries live session states, explains each change, and supports
rapid beat prototyping with an AI beat generation model and a Hum-to-MIDI model. DAWZY uses
LLM-based code generation as a novel way to significantly reduce interface familiarization time.

Track: Demo Track

Confirmation: Demo Track: I confirm that I have followed the formatting guideline and included all author names and affiliations.

(Optional) Short Video Recording Link: https://www.youtube.com/watch?v=RQmCuYLkEDk

Submission Number: 78