Is create-cartwright built on create-next-app?

No — it scaffolds a complete open-source Next.js engine (Cartwright) from a template repo. The result is a normal Next.js 16 project you could have built from create-next-app, with the product layers already implemented.

When is create-next-app the better choice?

When you are building something that is not a website, shop or content product — a dashboard, an internal tool, a novel app — and you want zero opinions in the repo. Blank is a feature there.

Can I strip Cartwright back to almost-blank?

Yes — it is your repo, and the engine is flag- and module-gated. But if you know you want near-blank, start with create-next-app instead; Cartwright earns its weight when you use the product layers.

Cartwright vs create-next-app

Both are one-command Next.js scaffolds — the difference is what exists when the command finishes. create-next-app hands you a blank canvas: routing and tooling, no pages, no data layer, no admin. create-cartwright hands you a running product on the same stack: a designed homepage, Prisma database with seeded content, an admin with auth, optional cart/checkout, JSON-LD, and an AI tool surface your coding agent can drive. Start blank when the app is genuinely novel and you want zero opinions; start with Cartwright when the goal is a website or shop — you skip the first two weeks of plumbing and still own every line, because it is your repo either way.

Dimension	Cartwright	create-next-app
What exists after the command	A running site: design system, seeded database, admin, auth, SEO/JSON-LD — and a webshop if you want it.	A blank app directory with routing, TypeScript and Tailwind configured. Everything else is yours to build.
Stack	Next.js 16 + React 19 + Tailwind v4 + Prisma + Stripe + NextAuth — pre-wired.	Next.js + your choices; official, minimal, unopinionated.
Freedom	Opinionated defaults you can rip out — it is a normal repo, not a framework on top.	Total — no opinions to remove. This is its genuine strength for novel apps.
AI-agent readiness	Agent rules files, an MCP server, a REST tool surface and plan-first admin tools ship in the box.	The 16.3 next-dev-loop skill helps agents iterate, but there is no product surface to drive.
Time to a shippable website/shop	Minutes (measured 99 s to a designed homepage with an AI agent).	Days to weeks — auth, data, admin, checkout and SEO are all still ahead of you.
Ownership	MIT; your repo from the first commit.	MIT; your repo from the first commit. Equal.

Frequently asked

Is create-cartwright built on create-next-app?: No — it scaffolds a complete open-source Next.js engine (Cartwright) from a template repo. The result is a normal Next.js 16 project you could have built from create-next-app, with the product layers already implemented.
When is create-next-app the better choice?: When you are building something that is not a website, shop or content product — a dashboard, an internal tool, a novel app — and you want zero opinions in the repo. Blank is a feature there.
Can I strip Cartwright back to almost-blank?: Yes — it is your repo, and the engine is flag- and module-gated. But if you know you want near-blank, start with create-next-app instead; Cartwright earns its weight when you use the product layers.

Try it in one command

npx create-cartwright@latest my-shop

MIT-licensed. Read the docs or see the full comparison hub.