我不再為每次微型發布都用 Next.js 重建登陸頁

I used to open a fresh Next.js app every time I had a small idea to validate.

Not because the idea needed React Server Components. Because I wanted a page that looked “real” — hero, pricing, FAQ, a place for an email — and Next was the hammer I already knew.

Then I’d burn two evenings on layout, fonts, a half-wired form, and a deploy that still felt like a mini product. By the time the landing was “good enough,” the urge to launch had cooled. The waitlist stayed empty. The Gumroad link never went live.

If that sounds familiar, this is the note I wish I’d written to myself a year ago.

The actual job of a launch page

For a micro-launch — a waitlist, a $19–$49 digital kit, a course teaser — the page has one job:

Get someone from “curious” to “email or payment” without making you rebuild a frontend repo.

That’s it. It does not need auth. It does not need a CMS. It does not need a design system with 40 tokens.

Heavy boilerplates (the $150–$300 “Ship your SaaS” class) are excellent when you already know you’re building a multi-tenant app. They are overkill when you only need a front door while the real product is still a weekend away.

Static HTML + Tailwind is boring on purpose. One folder. Edit copy. Host anywhere. Ship the same afternoon.

Why static HTML + Tailwind wins for waitlist / Gumroad launches

1. Time-to-first-visitor drops from days to hours

You clone or unzip, replace headlines and prices, drop in a Gumroad embed or a Tally/Formspree action, push to any static host. No npm run rabbit hole unless you want it.

2. The cost model matches the product

If you’re selling a one-time ZIP for $39, your acquisition page should not cost a week of engineering. Static hosting is free or near-free. No serverless form tax unless you choose a paid form tool later.

3. You can change the offer without touching an app

Swap “Join waitlist” for “Buy · $39.” Add an AI-tool vertical without forking a monorepo. Indie launches iterate on positioning more than on architecture in week one.

4. Buyers of digital kits understand “download and edit”

Your customer is often another maker. They don’t need a dashboard. They need a clean page they can restyle and redeploy. HTML is the universal export format.

I’m not saying never use Next. I’m saying: don’t start the Next project until the static page has proven anyone cares.

The structure of a page that converts (enough)

You don’t need twelve sections. You need a clear path:

Hero — who it’s for + the outcome in one breath. One primary CTA. Secondary can be “see demo” or “join waitlist,” not five competing buttons.
Benefits — three to five outcomes, not feature laundry lists. “Ship a pricing page today” beats “uses Tailwind v3.”
How it works — three short steps. Edit → host → collect emails or take payment. Reduce imagined complexity.
Social proof / honesty — if you don’t have testimonials yet, don’t invent them. Say you’re a solo maker, link a live demo, invite a roast. Trust compounds from honesty faster than from fake logos.
Pricing — one clear price. “$39 one-time” is easier to decide on than “starter / pro / agency” when you’re selling a kit.
FAQ — answer the objections you already hear: “Is this a SaaS boilerplate?” “Can I use my own form?” “Where do I host?”
Final CTA — repeat the one action. Buy, or waitlist. Not both equally loud.

Typography and contrast matter more than fancy motion. Big gray slogan text that fails WCAG contrast loses people who never scroll. Flat, dark, readable beats purple glow balls for a tool aimed at builders.

Deploy options that don’t eat the weekend

Anything that serves static files works:

GitHub Pages / Cloudflare Pages / Netlify / Vercel static — push a folder, get HTTPS.
A cheap VPS + nginx — fine if you already have one; overkill if you don’t.
Even a simple object store + CDN — if that’s your comfort zone.

Wire the form to Tally, Formspree, or your email tool. Wire the buy button to Gumroad or Lemon Squeezy. You’re done. The “backend” for a micro-launch is someone else’s form and checkout.

Save Next.js / full-stack boilerplates for the day after you have emails or sales and a reason to authenticate users.

What I built after getting tired of rebuilding

I packaged the pattern I kept rebuilding into LaunchShelf: a static HTML + Tailwind landing kit with hero, benefits, how-it-works, pricing, FAQ, waitlist slot, and a Gumroad/Lemon buy-button slot — plus short deploy notes.

Live demos (HTTPS, no signup wall):

General kit: https://ai.i258.cn/launchshelf/
AI-tool vertical copy: https://ai.i258.cn/launchshelf-ai/

If the demos are useful and you’d rather not assemble the same page again: $39 one-time on Gumroad — https://zhengchh.gumroad.com/l/launchshelf

No affiliate tricks. No “limited seats.” Just a template I wish I’d bought instead of rebuilding.

If you’re not buying yet but want vertical packs later (AI / course-waitlist / changelog-style), the waitlist is here: https://tally.so/r/2jj45e

No fake “10k makers trust us” line. I’m a solo maker. Early views, zero sales as of writing — which is exactly why I’m shipping the page and asking for signal instead of polishing another Next shell.

When you should not buy a kit

You already have a design system and a marketing site in production.
You’re launching a real multi-tenant SaaS next week and need auth, billing, and admin on day one — buy a serious boilerplate.
You enjoy building the landing as much as the product. That’s a valid hobby; skip the kit.

When static is enough

You’re validating a micro-tool, course, or digital download this month.
Your “site” is still a Notion doc or a Google Doc link in Discord.
You need emails or one payment before you write another layout.tsx.

Stop rebuilding the front door. Put something live, collect the signal, then earn the right to a heavier stack.

If you roast the demos in the comments — what’s unclear, what’s missing, which vertical you’d actually pay for — that’s more useful to me than a silent scroll. Thanks for reading.