Evidence

See ArchPilot work on real repositories.

We validate substantial open-source repositories so you can see the architecture ArchPilot discovers, the findings it raises, and the judgment needed to interpret them well.

TypeScript, NestJS, React, GraphQL, PostgreSQL

Twenty

A TypeScript CRM monorepo with a NestJS backend, React frontend packages, shared libraries, GraphQL, and PostgreSQL.

Health85
Readiness100
17
Components
16
Relationships

The run shows backend discovery in a large TypeScript monorepo, keeps test-only and scaffold-template packages outside default governed components, and separates two deterministic boundary findings in documentation tooling from warnings that require review.

View validation

TypeScript, NestJS, Angular, React, GraphQL

Vendure

A TypeScript commerce platform with NestJS backend packages, frontend packages, plugins, and shared libraries.

Health93
Readiness100
12
Components
8
Relationships

The run shows how ArchPilot turns a large commerce monorepo into a readable architecture inventory, highlights the core package as a dependency hub, and separates error-free validation from warnings that deserve review.

View validation

Java, Spring, Maven reactor

Spring Modulith

A Java and Spring multi-module repository with core libraries, event integrations, observability modules, examples, and integration tests.

Health88
Readiness100
31
Components
51
Relationships

The run shows nested module discovery, dependency hubs in shared event infrastructure, and scope-dependent event findings where examples and integration-test modules are intentionally governed.

View validation

Method

Simple enough to trust, specific enough to reproduce.

1

Pin the official commit

We record the repository URL, branch, commit SHA, and commit date before running ArchPilot.

2

Run the published CLI

We run each validation with a published ArchPilot CLI release and record the exact version with the repository evidence.

3

Publish explainable results

We show exact metrics only when the warnings and scope can be explained without implying endorsement.

How to read scores

Scores are context, not rankings.

Architecture Health reflects findings against the generated model. Setup Readiness reflects whether the repository has enough configuration and artifacts for governance. Warnings often require human review.

Open-source Architecture Validations | ArchPilot