Next.js starter kits for development agencies
Five developers, a dozen briefs a year, and the same three weeks of setup on each one.
What this is
An agency's problem is not that auth is hard. It is that five developers solve it five ways, and three projects later nobody remembers which one handles webhook replays correctly.
A starter kit is worth looking at as a shared default, not as a time-saver. The time saving is real but modest. The consistency is the thing.
Where the value actually is
| Without a shared starting point | With one | |
|---|---|---|
| Setup per project | 12–19 hours, varying by developer | 3–4 hours |
| Auth implementations in the codebase | One per developer | One |
| Onboarding a developer onto a project | Read this project's particular wiring | Read it once |
| Webhook correctness | Whoever wrote it that week | Verified and idempotent by default |
The last row is the one that costs money when it goes wrong. A webhook handler that is not idempotent looks fine until a processor replays an event and a customer is double-charged or double-provisioned. The kit's handlers key on the processor's own subscription id and upsert, so a replay updates rather than duplicating.
The licence, which is the part procurement will ask about
The terms permit unlimited personal and commercial projects and prohibit redistributing, reselling or publishing the source code as your own product. There is no per-seat tier and no renewal.
Two caveats worth raising before you standardise on it:
- The terms are written per project, not per developer. They do not set a seat limit, but they also do not say "any number of developers" in those words. If your procurement process needs that explicitly, ask for it in writing at hello@cheapstack.dev.
- If client contracts require full source assignment, raise that too. Building a client's application with the kit is commercial use; handing over the kit itself is closer to redistribution.
Neither is a reason not to buy. Both are reasons to spend one email getting it in writing rather than assuming.
What it does not solve
Being straight about this: a starter kit is a starting point, not a platform.
- It does not enforce conventions after week one. Your own review process does that.
- It does not cover your deployment pipeline, monitoring, or on-call.
- It will drift from upstream. You are buying a snapshot of source code, not a dependency you upgrade.
If you want something that stays current and has a support contract behind it, MakerKit at $349 is a more honest fit, and it is stronger on MFA, passkeys and i18n. The comparison page says so.
What to buy
The SaaS Boilerplate at $59, or the Pro Bundle at $79 for all eight kits. At agency rates the bundle is worth less than an hour of one developer's time.
Where to go next
- MakerKit vs CheapStack — the honest case for spending more
- Built with CheapStack — what has been built on these kits
- Documentation — what you are actually getting
Questions
Is there a per-seat or team licence?
There is no per-seat tier. The licence covers unlimited personal and commercial projects. Read it yourself at /terms before standardising a team on it.
Can every developer on the team use the same purchase?
The terms are written per project rather than per developer, and set no seat count. If you need that stated explicitly for a procurement process, email hello@cheapstack.dev and ask for it in writing.
Will a starter kit work as an internal standard?
Partly. It gives you a shared starting point for auth, payments and schema, which is the part everyone solves differently. It will not enforce conventions over time — the AGENTS.md and .cursorrules files in the kit do more for that than the code does.