Pieter Levels' Solo Business Approach

Pieter Levels is a Dutch indie maker who built a portfolio of profitable solo businesses including Nomad List, Remote OK, and Photo AI. He started with a “12 startups in 12 months” challenge in 2014 and has been shipping products from his laptop ever since, documenting everything publicly on X (@levelsio).

Pieter runs a $3.1M/year business alone. No employees, no funding, no complex tech stack. His approach is divisive — some call it genius, others call it reckless.

Here’s what I find interesting: his hit rate is around 5%. Only 4 out of 70+ projects made money. The successful ones just made enough to cover all the failures. This is the Small Bets philosophy in action.

The numbers

These are the winners. The graveyard of 66+ failed projects is invisible.

Vibe coding, for real

Levels takes vibe coding seriously. Accept AI suggestions fast, don’t overthink, fix issues when users report them.

This sounds terrible until you remember the context: most of his projects fail anyway. The goal is speed-to-validation, not code quality. If an idea works, you can always clean it up later. If it doesn’t, you saved yourself weeks of premature optimization.

“I need to go fast… within two weeks to see if idea works.”

The deliberately boring stack

This is the part that triggers people. His tech stack:

• Backend: Vanilla PHP. No framework. Single-file applications.
• Frontend: Plain HTML. jQuery.
• Database: SQLite. One file.
• Hosting: Single VPS on Linode.

No build process. No migrations. No Kubernetes. You can deploy with FTP if you need to.

Most developers find this horrifying. But the stack has survived years of real traffic. And when an AI writes your code, simple tech is easier to debug than complex abstractions you don’t fully understand.

The approach

1. Solve your own problems.

Photo AI: he wanted better profile pictures. Interior AI: wanted to redesign his apartment. Nomad List: needed places to live as a nomad.

The pattern: scratch your own itch. You already know the problem is real.

2. Charge money on day one.

Payment button before the product is polished. No free tier. If nobody pays, you know immediately. If they do, you have validation and cash flow.

3. Manual first, automate later.

Interior AI started with Pieter processing orders himself. Then scripts. Then cron jobs. Now it’s fully AI-operated with 99% margins.

The progression: do it manually → script the repetitive parts → automate the scripts → let AI handle edge cases.

4. Build in public.

His 600K+ followers on X watch him ship. They become early customers. The updates are simple: “Day 1: building X. Day 3: first customer. Day 7: $500 MRR.”

Revenue numbers, technical decisions, failures. Everything visible except customer data and API keys. He wrote about this approach in his book MAKE.

The shipping speed

fly.pieter.com: vibe coded in 3 hours. Launched on day 17. Now makes $87K/month.

That’s the cadence. MVP in a weekend, launch with Stripe attached, iterate based on what paying customers actually want. Kill projects that don’t get traction within two weeks.

The decision framework is simple:

• Less than 10 paying customers after 2 weeks? Kill it.
• Can’t get to $1K MRR in 30 days? Kill it.
• Hate working on it? Kill it.

If you’re getting traction and enjoying it, double down.

The margin game

Interior AI makes $40K/month with 99% margins. The math:

• Revenue per customer: $30/month
• AI API costs: $0.10-0.50 per customer
• Hosting: $20/month total
• Support: $0 (AI handles it)
• Employees: $0

Digital-only products, API-based delivery, AI customer service, single server. That’s how you get 99% margins as a solo operator.

What kills most indie projects

Building for months before launching. Free tiers that attract non-buyers. Complex tech that slows shipping. Perfectionism. Building what you think others want instead of solving your own problems.

The 5% hit rate is real. Plan for failure. Ship 20 projects to get 1 hit. Don’t tie your identity to any single project.

The honest timeline

Pieter’s path to $3M/year took 10+ years. First three years were mostly failures with small wins. Years 4-6 got him one product making $10K/month. Years 7-9 built a portfolio.

Your realistic year one: ship 3-5 projects, maybe 1 works. If you hit $1K MRR by month 6, you’re ahead of most.

What I think about this

The approach is polarizing. Ship fast, fix later, simple tech, high failure tolerance. It works for Pieter — the numbers are real.

Whether it works for you depends on your risk tolerance and how many failed projects you can stomach before one hits. Most people burn out before getting there. The ones who don’t either get lucky early or have enough runway to keep shipping.

The 5% hit rate is the honest part most people skip over.

Next: Build Your Personal AI OS