Skip to content

Model the product. Regenerate the software.

Turn product intent into software you can evolve, regenerate, validate, and deliver without making generated code a second source of truth.
One model · many verified projections

modelARch is an environment for discovering a product, modeling its domain and experience, and compiling that intent into a working application. The model remains authoritative; generated artifacts remain replaceable.

01

Discover

Explore the problem, actors, vocabulary, rules, exceptions, and unresolved decisions.

02

Design

Turn intent into journeys, views, interaction patterns, brand, and experience decisions.

03

Generate

Validate one model and compile it into application, tests, contracts, and delivery artifacts.

04

Operate

Promote immutable releases while data, configuration, infrastructure, and telemetry keep their own authority.

I want to understand modelARch

Learn the core concepts, the workflow, and what the platform produces. Start with Introduction and Mental model.

I want to model a product

Capture areas, objects, actors, actions, events, rules, queries, and integrations. Go to Modeling overview.

I need to evaluate the paradigm

Read its invariants, authority boundaries, migration strategy, release protocol, and incident model. Open Regenerative Software.

I need operational answers

Understand delivery channels, configuration, secrets, observability, rollback, and troubleshooting. Go to Operate & deliver.

AuthorityEvolutionary Modelrevision n
validate
CompilerGeneratorversion v
materialize
EvidenceReleaseartifacts + manifests
promote
Observed stateRuntimedata + config + infra

The short form is: the model evolves, the software regenerates, the data migrates, releases are promoted, and production is operated. Each kind of state has one explicit authority.

These docs deliberately separate three things:

  • The paradigm — the normative contract an implementation must satisfy.
  • The product — the modelARch experience for discovering, designing, generating, and delivering software.
  • Current support — concrete model constructs, generator targets, add-ons, and infrastructure capabilities.

That separation keeps the thesis honest: a principle can be robust before every supporting mechanism is complete, while a shipped capability must never be presented as merely aspirational.