How kUML compares

Feature	kUML	PlantUML	Mermaid	D2	Structurizr	MagicDraw	Modelio	Diagrams (Py)
UML 2.x (all diagram types)	✅	⚠️	❌	❌	❌	✅	✅	❌
SysML 2 first-class	✅	⚠️	❌	❌	❌	⚠️	⚠️	❌
C4 first-class	✅	⚠️	⚠️	❌	✅	❌	❌	⚠️
BPMN 2.0 first-class	✅	❌	❌	❌	❌	⚠️	⚠️	❌
ERM / DB modelling (4 notations)	✅	⚠️	⚠️	⚠️	❌	⚠️	⚠️	❌
Type-safe DSL	✅	❌	❌	❌	⚠️	❌	❌	⚠️
Real metamodel	✅	❌	❌	❌	✅	✅	✅	❌
OCL constraints	✅	❌	❌	❌	❌	✅	✅	❌
AUTOSAR profile (built-in)	✅	❌	❌	❌	❌	⚠️	❌	❌
Built-in UML profiles (5)	✅	❌	❌	❌	❌	⚠️	⚠️	❌
Code generation (M2T)	✅	❌	❌	❌	⚠️	✅	✅	❌
MDA transforms (M2M)	✅	❌	❌	❌	❌	✅	⚠️	❌
Executable behaviour simulation	✅	❌	❌	❌	❌	◐	❌	❌
M2M code generation (transformers)	✅	❌	❌	❌	❌	◐	◐	❌
Typed expression validation	✅	❌	❌	❌	❌	⚠️	❌	❌
Git-native	✅	✅	✅	✅	✅	❌	⚠️	✅
Open source	✅	✅	✅	✅	⚠️	❌	✅	✅
LLM-first design	✅	❌	❌	❌	❌	❌	❌	❌
CLI + Gradle + IDE + Web	✅	⚠️	⚠️	⚠️	⚠️	✅	✅	⚠️
Beautiful auto-layouts	✅	❌	⚠️	✅	✅	⚠️	⚠️	✅
Grid layout (opt-in, experimental)	✅	❌	❌	⚠️	❌	⚠️	⚠️	❌
M2M code generation (5 transformers)	✅	❌	❌	❌	❌	◐	◐	❌
Native installers (DEB/RPM/DMG/MSI/Docker)	✅	❌	❌	❌	❌	✅	◐	❌

Feature

kUML

PlantUML

Mermaid

Structurizr

MagicDraw

Modelio

Diagrams (Py)

UML 2.x (all diagram types)

✅

⚠️

❌

✅

❌

SysML 2 first-class

✅

⚠️

❌

⚠️

❌

C4 first-class

✅

⚠️

❌

✅

❌

⚠️

BPMN 2.0 first-class

✅

❌

⚠️

❌

ERM / DB modelling (4 notations)

✅

⚠️

❌

⚠️

❌

Type-safe DSL

✅

❌

⚠️

❌

⚠️

Real metamodel

✅

❌

✅

❌

OCL constraints

✅

❌

✅

❌

AUTOSAR profile (built-in)

✅

❌

⚠️

❌

Built-in UML profiles (5)

✅

❌

⚠️

❌

Code generation (M2T)

✅

❌

⚠️

✅

❌

MDA transforms (M2M)

✅

❌

✅

⚠️

❌

Executable behaviour simulation

✅

❌

◐

❌

M2M code generation (transformers)

✅

❌

◐

❌

Typed expression validation

✅

❌

⚠️

❌

Git-native

✅

❌

⚠️

✅

Open source

✅

⚠️

❌

✅

LLM-first design

✅

❌

CLI + Gradle + IDE + Web

✅

⚠️

✅

⚠️

Beautiful auto-layouts

✅

❌

⚠️

✅

⚠️

✅

Grid layout (opt-in, experimental)

✅

❌

⚠️

❌

⚠️

❌

M2M code generation (5 transformers)

✅

❌

◐

❌

Native installers (DEB/RPM/DMG/MSI/Docker)

✅

❌

✅

◐

❌

✅ present · ⚠️ limited or partial · ◐ proprietary/closed · ❌ absent

Where kUML wins

Modelling depth, Git-versionability and LLM friendliness in one tool. Nobody else combines all three.
Type-safe DSL as the model substrate — structurally superior for toolchain integration and AI-assisted generation.
UML, SysML 2, C4 and BPMN 2.0 as equal first-class languages. Every other tool forces you to pick one.
Apache 2.0 with a modern stack (Kotlin, Compose). No Eclipse legacy.
Designed for the LLM era from day one — structural, not accidental, fitness for AI-assisted modelling.

Where we acknowledge the competition

PlantUML’s ecosystem is enormous — wikis, IDEs, AsciiDoc tooling, CI plugins. kUML builds bridges (AsciiDoc extension, Gradle plugin, IDE plugins) rather than fighting on volume.
Structurizr owns C4 emotionally. kUML offers Structurizr DSL round-trip — interoperate, don’t compete.
Mermaid has GitHub-native rendering — the lowest possible barrier. kUML targets value from step two onward.
MagicDraw and Cameo dominate enterprise automotive — but their AUTOSAR support is a paid add-on. kUML ships the AUTOSAR profile in the box: «SoftwareComponent» with kind/packageName, «ComInterface» with version/isService, «AutosarPort» with direction (D17 — deliberately not Port to avoid the UML metaclass clash), «Runnable» with periodic/event-triggered/init/shutdown kinds, and «BehaviorSpec» on state machines feeding into kuml simulate for trace-based RTE verification.