Composing at five levels

Spatial composition is not one activity but a stack of activities. A creator can work at the tick level, the topology level, the agent level, the modifier-chain level, or the song-structure level, and most interesting compositions move fluidly between several. Knowing which level you are operating at is half of knowing what to do next.

When a creator opens a blank session in Vibe Together, they are facing a stack of compositional choices that operate at different scales of time and structure. The platform supports all of them at once, but you make better decisions when you are clear about which one you are working on. Here are the five levels, from the smallest grain to the longest form.

Level 1: The tick

The smallest unit. A tick is one step of the platform clock: at the default 120 BPM with TICK_SUBDIVISION = 4, that is 125 milliseconds, a sixteenth note. Every signal arrives, fires, and travels in tick-quantized increments. When you hear a precise rhythm or a tight delay, you are hearing tick-level decisions.

Moves at this level include setting a Source's interval trigger, pinning a wire to an exact millisecond delay, or tuning a Chord Node's strum length so it spreads across the intended ticks.

Level 2: The topology

The shape of the graph itself: where nodes are placed, which wires connect what, what the layout looks like. This is the level where layout-as-rhythm lives: moving a node changes the music because the topology is the score.

Moves at this level include dragging a node farther from its Source, branching one Source into parallel voices, or deciding whether a feedback path should exist. Topology is tangible: you compose by placing.

Level 3: The agent

Inside every Source is a 30-parameter procedural agent that decides what notes to generate. The agent operates below topology: it makes per-signal choices about pitch, voicing, contour, dynamics, articulation, anticipation, and behavior. Two Sources with identical topology but different agents produce wildly different music.

Compositional moves at this level: choosing a preset (bass, melody, harmony, rhythm, lead, percussion), or opening the custom preset panel and editing the contour from free to gravity, the harmonic focus from scaleTones to guideTones, the swing amount from 0 to 0.33. This is the level where spatial composition feels closest to traditional procedural music.

The full 30-parameter reference for the agent. Continue learning

Level 4: The modifier chain

Between a Source and its destinations, Splitters, Gates, Harmony nodes, and Modulators can transform every signal in flight. This is where behavior gets composed: probability, timing, transposition, and role-swapping. The endpoints can stay the same while the chain between them makes very different music.

Compositional moves at this level: inserting a splitter to introduce variation, adding a gate to reshape the rhythm, threading a harmony chain to build counterpoint from a single voice, dropping in a modulator to change the personality of a branch mid-flow. This is where spatial composition shows its most distinctive expressive range: there is no equivalent in traditional notation.

Level 5: Song structure

Above the graph is song structure: named sections, per-section progressions, and section-aware Source triggers (sectionStart, sectionEnd, sectionCycle). This is where form gets composed.

Moves at this level include defining section lengths, assigning local progressions, and using section triggers so certain Sources only fire on transitions. The same agent, topology, and modifier chain can behave differently in a verse and a chorus.

How to enable and configure song structure. Continue learning

Why this matters

Most musical traditions privilege one level over the others. A pop songwriter usually thinks in song structure first and the smaller scales second. A drum programmer usually thinks at the tick level. A jazz comper usually thinks at the agent level: what notes are appropriate against this chord, and lets topology emerge. Spatial composition lets a creator move between levels by choice, picking the level that best matches the question they are trying to answer.

Collaborators can work at different levels without leaving the shared object. One person tunes a melody Source, another rearranges topology, and a third shapes song structure. The graph keeps those decisions in one place.