Locara

Pieter Levels (@levelsio)

Who: Solo Dutch indie maker, founder of Nomad List, Remote OK, Photo AI, Interior AI, and ~10+ other products. Probably the best-documented case of high-revenue solo build-in-public. Status: Multi-million dollar revenue, no employees, no investors, fully transparent revenue dashboards. Author of Make — a playbook for indie product building. Most relevant to Locara: Operational template for “indie OSS founder ships meaningful product at scale without VC.” Also the cleanest example of build-in-public actually working as a growth engine.

Background

Levels started in 2014 with “12 startups in 12 months” — a public commitment to launch one product per month. Most flopped, two (Nomad List, Remote OK) became durable businesses. By 2024 he’d added Photo AI and Interior AI (both AI-image-gen wrappers, both very profitable). His public revenue dashboards (over $3M ARR across products) became part of the brand.

He’s an outlier, not a model — most builders following the playbook fail. But the principles are real and transferable.

Key principles / decisions

  • One person. No employees, ever. Hires contractors for narrow tasks (legal, design) but keeps headcount at one.
  • No investors. Never raised. Bootstrapped from product 1.
  • Public revenue. Real-time dashboards. Marketing-as-honesty.
  • PHP + jQuery + SQLite as the stack. He’s vocal about it. Boring tech, fast to ship, runs on a single VPS.
  • Stripe Atlas for company structure (Delaware C-corp despite being non-US resident).
  • Ship every day. Public commits, public progress.
  • Audience as moat. ~700k+ Twitter followers. Distribution is the actual product when you’re solo.
  • Anti-Twitter-thread, pro-Twitter. Long-form blog posts on his site, short banger tweets for distribution. Clear separation.
  • No “investors deck” or “go-to-market plan.” Just: build, launch, iterate, charge from day one.
  • Strong opinions on tooling. Anti-microservices, anti-AWS-Lambda, pro-monolith, pro-VPS, pro-PostgreSQL-or-SQLite, anti-React-for-CRUD-apps.
  • Shipping is healing. Levels has been explicit that the shipping cadence is psychologically necessary for him.

What worked

  • Transparency built trust faster than marketing could. Open MRR + public roadmap + open struggles → high signal.
  • Audience-first distribution. By the time Photo AI launched, he had 500k+ followers; “free” launch traffic is enormous.
  • Cheap infrastructure stack. SQLite + PHP + single server keeps OpEx low; profit margins are 90%+.
  • Multiple products diversify. Each product earns enough; no single one is do-or-die. Risk spread without team coordination cost.
  • Boring tech keeps him shipping. No dependency hell, no microservice debugging, no Kubernetes. Time goes into product.
  • The Make book turned the philosophy into a repeatable artifact others reference.

What failed / criticisms

  • Survivorship bias is severe. “12 startups in 12 months” → 2 hits. Most followers of the playbook ship 12 failures.
  • Mental health cost has been public. He’s open about the loneliness, the burnout cycles, the stress of solo accountability. Not a sustainable model for everyone.
  • Quality bar is “good enough,” not great. Products often look amateur, have rough edges, lack polish. Works for his audience, fails in many segments.
  • Privacy / trust occasionally an issue. His sites have been hit with controversies over moderation, data handling. Fewer eyes on the codebase = fewer reviewers.
  • Brand is heavily personal. If something happens to him, the products are at significant risk.
  • PHP + jQuery is genuinely worse than modern alternatives for some classes of app. He optimizes for shipping speed, not technical excellence.
  • Anti-investor stance reads as ideological to some founders — sometimes capital genuinely is the answer.

Specific learnings for Locara

  1. Build-in-public works when the audience is distinct from the customer base. Levels’ audience is other indie devs; his customers are mostly different. Locara’s audience would similarly be “people interested in local AI” — devs and prosumers — and the real-time-progress story is content for them.
  2. Public commits + public roadmap + public ADRs from day one. Like Levels’ stack but for an OSS infrastructure project. Steal the format.
  3. Boring stack for the platform infra. Locara’s registry/website should run on dead-simple infra (static site + small API + Postgres on a managed host). Don’t build the “cool kid” stack for what is fundamentally a CRUD-and-CDN app.
  4. Audience-first distribution. Locara’s growth comes from the same kind of slow audience-building Levels did. Newsletter + Twitter + technical posts. Don’t pay for ads.
  5. Charge from day one if charging at all. If Locara ever does monetize anything (e.g., a “verified privacy” certification), charge from launch. Free → paid migrations are brutal.
  6. Multiple products as risk spread — Locara plans transcribe + DocVault as the first apps. This is structurally similar to “12 startups in 12 months” but more focused. Ship two, see which resonates, double down.
  7. Single-person bandwidth caps mean ruthless prioritization. Levels’ philosophy: “if it’s not on fire, don’t fix it.” Adopt this. The 80/20 ratio is real.
  8. Personal brand risk. Don’t make Locara only a personal project. Plan for it to outlive your active involvement: clear license, contribution structure, runbooks.
  9. Privacy / trust failures are reputation risks. Levels has been hit; Locara would be hit harder because the entire pitch is privacy. Invest in security from day one.
  10. The Make book is worth reading. Cheap, fast, dense with patterns.

References