Overview

MTP (Mapping the Prompt) is a framework for steering LLM output through structured controls — sliders, grid coordinates, and presets — instead of long natural-language behavior instructions.

It controls model tone and style through a coordinate-based compilation step rather than direct verbal instructions. Instead of relying on requests such as “be concise” or “think deeply,” it uses a position in a structured space: a slider value, a grid coordinate, or a named preset. The compiler deterministically converts that position into constraint XML that reshapes the model’s output.

This page summarizes core terms and a lightweight description of the two layers behind MTP’s controls. In practical terms, those layers are grid position and intensity / polarity; in the formal model they are called Space and Intensity. Full equations and zone boundaries are covered in Grid and Coordinate System; per-node behavior is summarized in Node Reference.

---

Core terms

The following terms recur throughout MTP documentation. Together they describe the pipeline from user input to applied constraints:

  ┌─────────┐     ┌──────────────────┐     ┌───────────────────────────┐     ┌─────────────┐
  │  Input  │ ──→ │     Compiler     │ ──→ │  Constraint extraction    │ ──→ │   Output    │
  └─────────┘     └──────────────────┘     └───────────────────────────┘     └─────────────┘

  preset ─── expands to ──→ grid coordinates ──┐
  slider (node:intensity) ───────────────────→ ├→ (axis, polarity, intensity) → node file → constraints
  grid coordinate (column:row) ──────────────→ ┘

Term	Meaning
Node	One of the nine semantic axes (for example, Red). Each node has a Side A name (for example, Power) and a Side B name (for example, Void).
Axis	The color-based identity used to distinguish nodes: Yellow, Red, Magenta, Green, Transparent, White, Cyan, Blue, Purple.
Side A / Side B	The two poles of each axis. Positive intensity activates Side A; negative intensity activates Side B.
Intensity	A strength value from 0 to 100. It determines which constraint tier (Low / Mid / High) is extracted.
Slider	Input in node:intensity format. Examples: power:70, void:80, yellow:-30
Grid	Input in column:row format on the 19×19 grid. Examples: D:16, A:1. The compiler calculates axis, polarity, and intensity from the position.
Zone	One of the nine large cells created by dividing the grid into 3×3. Each zone maps to one axis.
Preset	A named blend that expands into grid coordinates before parsing. Example: strategist expands to P:16 P:4.
Constraint	Instructions extracted from a node file according to the resolved intensity tier. These instructions shape output behavior.

All input paths ultimately converge on the same internal representation: (axis, polarity, intensity). Inputs such as power:100, red:100, and J:4 resolve to the same three components and therefore extract the same constraints.

---

Two layers

At a formal level, MTP consists of two layers: Space and Intensity. This overview introduces them in concrete terms as grid position and intensity / polarity. Space is where something sits on the grid: axis and zone. Intensity is how strong it is and which pole is active; resolving grid coordinates involves Volcano mapping and tier extraction. The design background behind the 3×3 layout is covered in Design Background.

Layers at a glance

Layer	Controls	Key idea
Space	Grid position: which axis and zone	19×19 grid, 3×3 macro zones, hue cycle
Intensity	Intensity / polarity: strength and active pole	0–100 slider, Side A / Side B, strength and polarity mapped from distance to center

For each token, the compiler resolves (axis, polarity, intensity) and extracts tiered constraints from node files.

Space (grid position)

The interface uses an integer coordinate system based on 19×19 line intersections, which is the current default implementation. The surface is divided into nine macro zones in a symmetric 3×3 layout, and each zone maps to one color axis. The design rationale behind that arrangement is covered in Design Background.

Intensity (strength / polarity)

Each node has a continuous strength value from 0 to 100. Polarity determines whether Side A (positive) or Side B (negative / inverted pole) is active. Macro-zone color names can also be used as slider aliases (for example, yellow:50 ≡ open:50 on the Yellow axis). Strength and polarity from a grid position are computed with a Chebyshev-based radial rule, defined in Grid and Coordinate System. Qualitatively: neutral at the center, Side A peaks on the surrounding ring, Side B on the outer frame.

When several MTP tokens are present, the compiler emits one <constraints> block per token in parse order (not a separate layer). Optional design readings of token order (for example trajectory metaphors) are in Design Background; per-node summaries are in Node Reference.