MusiCue Chisels Music Into Frame-Perfect Animation Cues

MusiCue is an open-source tool that converts songs into typed, timeline-based cues for driving animation and show-control software. Developer cedarconnor built it to break audio into separated stems, beats, drum classifications, vocal phrases, and continuous energy curves. The three-layer pipeline runs locally on a Windows PC with an NVIDIA GPU and exports to nine formats — including After Effects, TouchDesigner, MIDI, and Unreal Engine.
Cedarconnor created MusiCue to give motion designers and live show operators a precise, structured alternative to simple amplitude-based reactivity. The analysis–compile–export architecture keeps the data customizable and repeatable. Because everything stays on your own machine, it avoids subscription fees and keeps sensitive media private.
Stem separation and beat detection included
- Separates audio into drums, bass, vocals, other.
- Detects beats, downbeats, sections, and tempo changes.
- Classifies drum hits by type (kick, snare, hi‑hat).
- Phrase detection for vocal and melodic lines.
- Four preset grammars for concert, animation, lighting, editing.
- Exports to After Effects, TouchDesigner, MIDI, Unreal, and more.
MusiCue suits anyone who needs music‑synchronized visuals but wants to work offline. Motion graphics artists, lighting programmers, and live‑show creators can generate precise timing data directly from their audio files. Running fully on a local NVIDIA GPU keeps projects private and avoids per‑seat cloud costs.
Developer notes and known limitations
The project is in active development, with a complete analysis backbone and a web UI that still requires a manual front‑end build. Some optional machine‑learning components, like All‑In‑One section detection, can be tricky to install on Windows and may not appear automatically. The developer plans to add audio export in a future update, while the tool’s grammar system makes it straightforward to add new target formats.
“I built a tool that turns a song into keyframes.” — Source: Reddit