wtg2

WardleyToGo

A Go library and tools for creating and rendering Wardley Maps.

A map is a directed graph of components, each with a position on a 100x100 coordinate system representing visibility (Y) and evolution (X). The entry point of this API is the Map structure.

The project maps itself using WTG2 — read the full strategic analysis or see the source.

Quickstart

WTG2 is a human-friendly DSL for describing Wardley Maps. wtg2svg reads WTG2 from stdin and writes SVG to stdout. See the full grammar for details.

Get wtg2svg

Download a binary from GitHub Releases, or install with Go:

go install github.com/owulveryck/wardleyToGo/cmd/wtg2svg@latest

Usage

Create a file hello.wtg2:

title: My First Map
anchor User : III.5
Platform : II.5
Data : III.8 (buy)
User -> Platform -> Data

Then render it:

wtg2svg < hello.wtg2 > map.svg

Try it without installing

Use the online playground to write and render WTG2 maps directly in your browser.

Generate maps with AI

The file wtg2/skill.md is a ready-to-use prompt that teaches any LLM (ChatGPT, Claude, Gemini, etc.) to generate valid WTG2 maps. Copy-paste it as a system prompt or into a conversation, then ask the model to create a Wardley Map for your domain — it will output WTG2 that you can render with wtg2svg or the playground.

WTG2: the Wardley Map DSL

WTG2 is a human-friendly domain-specific language for designing Wardley Maps. The full grammar is in wtg2/grammar.bnf.

Example

title: Plateforme de Navigation -- Strategie 2026
stages: Genesis, Custom, Product, Commodity

anchor User
anchor Partner

Application : III.5
Routing Engine : II.7 !! >> III.5 {
  type: build
  color: #3498DB
  note: "Key differentiator"
}
Map Data : III.1 (buy)
Cloud Infra : IV.3 (buy)

pipeline Routing Engine {
  Classic Algo : III.5
  Predictive AI : II.3
}

User -> Application -> Routing Engine -> Map Data
Application -> Cloud Infra
Routing Engine -> Cloud Infra

note "Outsource candidate Q4" on Cloud Infra
warning "Single vendor dependency" on Map Data
signal accelerating on Predictive AI

Language features

Feature	Syntax
Metadata	title:, date:, author:, scope:, question:
Custom stages	stages: Genesis, Custom, Product, Commodity
Anchors	anchor Name
Components	Name : III.5 or component Name : III.5
Types	(build), (buy), (outsource)
Block config	Name : II.7 { type: build, color: #3498DB, note: "..." }
Evolution	II.7 >> III.5 (with optional inertia !!)
Pipelines	pipeline Name { Member1 : III.5 }
Edges	A -> B -> C (chains supported)
Annotated edges	A -[label text]-> B
Submaps	submap Name : III.6
Groups	group Name { ... } (visual only)
Annotations	note "text" on Name, warning "text" on Name
Signals	signal accelerating|stagnating|declining on Name
Comments	// line or /* block */

Position notation

Positions use roman numerals (I-IV) for evolution phases, with a decimal for sub-positioning:

• I.0 = far left (genesis), IV.9 = far right (commodity)
• Each phase spans 25% of the axis
• Without decimal (e.g. III), the center of the phase is used

Using the library

Creating a map programmatically

First, create a component type:

type dummyComponent struct {
	id       int64
	position image.Point
}

func (d *dummyComponent) GetPosition() image.Point { return d.position }

func (d *dummyComponent) String() string { return strconv.FormatInt(d.id, 10) }

func (d *dummyComponent) ID() int64 { return d.id }

Then a collaboration structure (an edge):

type dummyCollaboration struct{ simple.Edge }

func (d *dummyCollaboration) GetType() wardleyToGo.EdgeType { return 0 }

func (d *dummyCollaboration) Draw(dst draw.Image, r image.Rectangle, src image.Image, sp image.Point) {
	coordsF := utils.CalcCoords(d.F.(wardleyToGo.Component).GetPosition(), r)
	coordsT := utils.CalcCoords(d.T.(wardleyToGo.Component).GetPosition(), r)
	drawing.Line(dst, coordsF.X, coordsF.Y, coordsT.X, coordsT.Y, color.Gray{Y: 128}, [2]int{})
}

And finally create the map: