I Built a Nuxt 4 Starter Template I'd Actually Use in Production

Maxime Bignolet — Thu, 05 Feb 2026 08:30:31 +0000

After a year of building Nuxt apps professionally at a French tech company, I noticed I was repeating the same setup for every project. So I built a starter template that I actually use in real work—and I'm sharing it today.

🌐 Live Demo • 📦 GitHub

The Problem with Starter Templates

Most templates fall into two categories:

Too minimal – Just the framework, you configure everything yourself
Too bloated – 50 dependencies, complex folder structures, half you'll delete anyway

What I needed was the middle ground: production-ready without the bloat.

What's Inside

✅ Nuxt 4 (latest stable)
✅ TypeScript (strict mode enabled)
✅ TailwindCSS 4 (yes, already!)
✅ Vite (lightning-fast HMR)
✅ ESLint (configured but not annoying)
✅ Clean folder structure

Clean Folder Structure

app/
├── assets/          # Global styles, images
├── components/      # Vue components (auto-imported)
├── composables/     # Reusable logic
├── layouts/         # App layouts
├── pages/           # File-based routing
└── utils/           # Helper functions

No src/ directory. Nuxt 4's new structure is cleaner without it.

ESLint Configuration That Works

Most templates either skip ESLint entirely or configure it so strictly that you fight with it constantly.

My sweet spot:

// eslint.config.mjs
export default [
  ...nuxtEslintConfig,
  {
    rules: {
      'vue/multi-word-component-names': 'off', // Too annoying
      '@typescript-eslint/no-unused-vars': 'warn', // Warning, not error
    }
  }
]

Catches real bugs without yelling about style preferences.

Real-World Numbers

From actual production use:

Metric	Result
⚡ Dev server start	~2 seconds
📦 Build time	~30 seconds
💾 Bundle size	~85KB gzipped

What I've Built With It

I've used this exact template to bootstrap:

🚀 A SaaS dashboard (added auth + Stripe)
📄 A marketing website (added Nuxt Content)
🔧 An internal tool (added API integration)

Each time: working, type-safe, styled app in under an hour.

Quick Start

Use GitHub's Template Feature (Recommended)

Click "Use this template" on the GitHub repo

Or Clone Manually

git clone https://github.com/MaximeBignolet/nuxt-starter-template.git
cd nuxt-starter-template
npm install
npm run dev

Visit http://localhost:3000 and you're ready to build 🚀

Key Lessons Learned

1. Lock Critical Dependencies

{
  "dependencies": {
    "nuxt": "^4.0.0",  // ✅ Minor updates OK
    "vue": "3.5.13"     // ❌ Locked - breaking changes happen
  }
}

Learned this the hard way when a Vue patch broke SSR in production.

2. Document Auto-Imports

Nuxt auto-imports are magical... until a new dev can't find where useAsyncData is defined.

Added a simple comment:

Auto-imported:
- Components from ~/components
- Composables from ~/composables  
- Utils from ~/utils

3. Examples > Documentation

Instead of documenting every config option, I included working examples:

✅ Page with async data fetching

✅ Reusable API call composable

✅ Layout with header/footer

✅ Component using Tailwind utilities

Delete what you don't need. Build on what you do.

Roadmap

Actively maintaining based on real project needs:

[ ] Dark mode toggle component
[ ] SEO meta management setup
[ ] i18n configuration example
[ ] Testing setup (Vitest + Playwright)

Not planning to add:

CMS integration (too opinionated)
Backend setup (everyone's different)
Complex state management (add if needed)

Try It Yourself

The template is MIT licensed and completely free to use:

🌐 Live Demo

📦 GitHub Repository

⭐ Give it a star if you find it useful!

Final Thoughts

The best starter template is the one you'd actually use. Not the one with the most GitHub stars or the most features—the one that gets you from idea to working app the fastest.

If you're starting a new Nuxt project and want something production-ready without the bloat, give it a try.

Discussion

What's your experience with starter templates? Do you prefer minimal setups or feature-rich boilerplates? What features would you add to this template?

Drop a comment below! 👇

Connect with me:

🌐 Portfolio
💼 LinkedIn
🐙 GitHub

Happy coding! 🚀

I built a tool to spot real product pain without reading forums for hours

Maxime Bignolet — Wed, 14 Jan 2026 18:06:35 +0000

Like a lot of founders, I kept starting projects from vague ideas, Twitter vibes, or half-remembered complaints.

The problem wasn’t a lack of ideas — it was signal extraction.

Reading forums manually is slow, noisy, and mentally draining:

- 90% of posts are advice, opinions, or meta discussions

- the real pain is buried in a few frustrated sentences

- after 30 minutes, you don’t know what actually repeats

So I built I Wish There Was.

It scans recent public forum discussions indexed by search engines, filters for explicit everyday frustrations (time wasted, confusion, friction, costs), and clusters them into clear problem statements.

For each cluster, it outputs:

- a concise problem title

- a short summary

- who is affected

- why it hurts

- verbatim user quotes

- a recurrence level (low / medium / high)

Important:
This is not market validation.
It doesn’t tell you what to build or guarantee demand.

It’s a signal tool — meant to help you decide what’s worth exploring next, instead of starting from a blank page.

The UX is intentionally simple:

- pick topics (and optional keywords/sources)

- choose a time range (7 / 14 / 30 days)

- run a scan

- get results in minutes

Pricing is credit-based (1 free scan, then small one-shot packs).
No subscription — this is a tool you use at a decision moment, not every day.

I’m sharing this here to get feedback from builders:

- Does this feel useful in your workflow?

- What would make the output more actionable without turning it into “fake validation”?

- Where would you personally use something like this?

https://i-wish-there-was.com/

Happy to answer questions or explain how it works under the hood.

Forem: Maxime Bignolet