Code agent controls digital audio workstation for precise mixing
TOMORI: A Tool-Orchestrated Mixing Operator for REAPER Interaction
Sound
Summary
Mixing music often involves vague descriptions of how a final sound should be, but it’s hard to know exactly what edits are needed to get there. The authors developed TOMORI, a system that uses code to directly control a popular music editing software called REAPER. Their system inspects projects, measures audio, and runs commands generated by AI to make mixing changes. They tested it with many examples and found that clear instructions work well, but vague requests are harder to achieve exactly, revealing challenges in translating creative desires into precise software actions.
What this means in practice
- •For audio software developers: Integrate code-driven control into DAWs to automate mixing tasks based on explicit user instructions.
- •For music production teams: Use explicit scripted commands to reliably reproduce reference audio mixing decisions within REAPER projects.
Authors
Yining Wang, Tianjiao Li
Abstract
Code-based agents can operate creative software, but mixing requests often describe an intended sound without specifying the edits needed to achieve it. We present TOMORI, a code-based mixing-agent harness for native digital audio workstation (DAW) control. Custom plugins provide REAPER project inspection, optional audio measurement, and execution of model-generated Lua code. All observations return as text; delivery includes an editable project and rendered audio. TOMORI-BENCH pairs 110 human-reviewed reference edits with three instruction forms and evaluates native DAW execution end to end through operation structure, parameter precision, and rendered-audio outcomes. Across 990 cases from three primary models, explicit instructions yield high reference agreement, while abstract requests reduce project agreement and audio improvement. Paired ablations show that the full workflow improves explicit execution over a single submission. Prompted audio measurement succeeds in all 28 selected abstract cases but yields no clear overall improvement on the reference-based metrics and increases tool calls. These results expose the gap between executable DAW control and precise realization of perceptual mixing goals.