Next.js SaaS starter + shadcn/ui: compatible, not locked

The question nobody in this search result answers

Search for a Next.js SaaS starter with shadcn and you get lists. Which kits use shadcn, how many stars they have, what they cost. Fine. But that is not the thing you are actually nervous about.

The real worry is narrower. You chose shadcn deliberately — you wanted to own the components, paste them from a registry, and let your editor's MCP pull new ones on demand. Then you consider paying $149 for a starter, and the obvious risk is that the kit has quietly built its own UI layer on top, so the registry, the CLI and the MCP stop being useful the day after you buy. That is why the free official starter feels like the safe default: it is canonical shadcn, and nothing sits between you and the registry.

That fear is reasonable. It is also testable in about sixty seconds, which is what the rest of this page is about. We ran the test on our own repo and published what came out, including the part that did not work at first.

What "shadcn-compatible" means here, precisely

Let's be exact, because this is the kind of claim buyers verify thirty seconds after purchase and request a refund over. Three separate statements, all true:

The sixty-second test, and its actual output

Here is the whole proof. Clone the repo, then:

$ npx shadcn@latest add dialog

✔ Checking registry.
✔ Installing dependencies.
✔ Created 1 file:
  - src/components/ui/dialog.tsx

The interesting part is not that it succeeded. It is the top of the file it wrote:

import { Dialog as DialogPrimitive } from "radix-ui"
import { cn } from "@/lib/utils"
import { Button } from "@/components/ui/button"

A stock registry component resolved our hand-rolled cn() and composed our hand-rolled Button. No edits. That is the whole argument for the token contract: the boundary between "our code" and "registry code" is just CSS variables, so the two sides compose instead of fighting.

We also checked that the theme genuinely lands rather than assuming it. Compiling the stylesheet and grepping the output, all seventeen utilities the dialog relies on resolve to our variables:

.bg-popover     { background-color: var(--popover); }
.bg-destructive { background-color: var(--destructive); }

One honest consequence. Because we do not vendor Radix, adding a registry component that needs it will pull radix-ui into your dependency tree at that moment. We think that is the right default — you get primitives when you ask for them, not before. If you would rather every primitive be present on day one, the official starter already does that, and that is a legitimate reason to prefer it.

The matrix, including the free one that beats us on price

Most comparison tables on paid-starter sites quietly omit nextjs/saas-starter, which is the single most likely thing you are comparing against. Ours used to as well. It is here now, in the first column, because a comparison that hides the obvious alternative is not a comparison.

The last row is the one no competitor publishes: when each kit last shipped a commit. For closed-source kits that number is simply not verifiable from outside, and we say so rather than guessing.

Official starter figures read from the GitHub API and package.json on main, 16 July 2026. Stars and commit dates move; re-check them before you decide. The "typical paid kit" column is deliberately vague because we will not publish specific numbers about closed-source competitors we cannot verify.

$0 vs $149: what the official starter makes you build yourself

The free starter's README does not oversell itself. It says, verbatim:

"While this template is intentionally minimal and to be used as a learning resource, there are other paid versions in the community which are more full-featured."

— nextjs/saas-starter README

We are not going to argue with that. It is an excellent piece of code and it does exactly what it says. It gets you auth, Stripe, Drizzle and real shadcn/ui for nothing. The honest pitch for paying us is not that it is bad — it is that it deliberately stops at the point where the boring six weeks begin. Here is that delta, with the time it actually took us to build each piece:

That is roughly 69 hours, and that is our time on a stack we wrote and know by heart. Your number will be higher on unfamiliar code — assume something like 90 to 100 hours, and more if you have not wired Stripe webhooks before. At any freelance rate you would recognise, that is the entire argument. If your hourly rate is above about $2, $149 is cheaper than building it.

The inverse is also true and worth saying plainly: if those seven rows are things you do not need, the free starter is the correct choice and you should close this tab. A blog and eight locales are not worth $149 to someone validating an idea over a weekend.

Stack currency, stated plainly

We run Next.js 15.5.14 and React 19.2.4. Not Next.js 16. If a comparison page tells you the free official starter is on Next.js 16 and we are behind, check its package.json yourself — it pins next@15.6.0-canary.59. Both kits are on Next.js 15. Theirs is a canary build, ours is a stable release, and neither of those facts is worth much marketing noise.

Our position on Next.js 16: we will move when the ecosystem does, not before. A starter's job is to be boring on release day. Being one stable minor behind and fully green beats being on a canary that breaks Clerk or Prisma in a way you get to debug on our behalf.

The freshness number that actually matters is the last row of the matrix above. The free starter's most recent commit, as of 16 July 2026, is from 11 December 2025 — and it was a security bump (CVE-2025-55182). That is not an attack on it; unmaintained is the normal state of a repo whose stated job is teaching. But if you adopt it, understand that keeping it current is now your job, and that is a real cost you should price in next to $149.

What's not included, and who should not buy this

Every kit in this search result lists what it has. None list what they lack, which is the cheaper and more useful list. Ours:

If you are learning how auth and billing work, do not buy this — you would be paying to skip the education you are after. Clone the free one. This kit is for people who have already built this plumbing at least once, know exactly how long it takes, and would rather not do it a third time.

Questions people actually ask