I built 4 SaaS platforms in 5 weeks. Here's everything.

No team. No funding. No agency. Just one person, a VPS, and a stupid amount of late nights. This is the honest breakdown — the numbers, the products, how it's possible, and what it means for what you can build right now.

I'm Eric. I run Cornerstone Digital out of Poole, Dorset. Five weeks ago I had one product live — a half-finished health coaching app. Today I have four production SaaS platforms running on a single VPS, every one of them with real users, real Stripe integration and real AI features baked into the core. All of it built solo. None of it outsourced.

Here's the part that matters for you reading this: I'm not special. I'm a one-person shop in a coastal town in the south of England. What I am is someone who figured out the new tooling early and went hard. The work below isn't a flex — it's a worked example of what a single person can ship in 2026 if they stop overthinking and start building.

Let me show you the four platforms, the actual numbers, and then I'll tell you how it's possible.

The four platforms

That's 518 source files, 157 API routes, 149 pages, 81,447 lines of production code across four separate codebases. Every line was written, reviewed, debugged and deployed by one person. Every Stripe integration, every database migration, every AI prompt, every Caddy config, every PM2 process — all of it me.

"That's not a side project. That's not a portfolio. That's a real working business that exists because one person decided to build it."

How is this even possible?

I'll be honest with you, because the alternative is pretending this took some kind of superhuman effort, and that's not true and it's not useful. Here's the real answer:

1. AI tooling is genuinely a 10x change for solo builders

I'm not talking about ChatGPT writing my marketing copy. I'm talking about pair-programming with Claude through every architectural decision, every database schema, every tricky bug. The right model used the right way doesn't replace the developer — it removes the friction that used to slow them down. The bottleneck used to be typing speed, lookup time and debugging dead ends. None of those are bottlenecks anymore.

If you're a developer reading this and you've tried AI coding tools and bounced off, try again. The difference between using them badly and using them well is enormous, and the gap between one and the other is closing every month.

2. I picked one stack and stopped second-guessing it

Every one of these four platforms runs on the same stack: Next.js 14 App Router, TypeScript, Supabase (Postgres + Auth + RLS), Anthropic Claude API, Stripe, deployed with PM2 behind Caddy on a single Ubuntu VPS. That's it. No microservices, no Kubernetes, no separate auth provider, no managed database per app, no serverless lambda spaghetti.

Picking the stack once means I never burn time on setup. A new project starts with a 90% solved infrastructure problem. I can have a new product deployed at app.cornerstonedigital.co.uk in under an hour because the path is muscle memory.

3. I build for myself first, sell second

Three of the four platforms started as scratch-my-own-itch builds. Dinara exists because my partner has a real, complicated set of health needs that no app on the market handled properly. MyTradeSite exists because I watched local tradesmen pay £200/month for terrible WordPress sites built by agencies who never picked up the phone. AgentAI exists because I noticed the UK estate agency market is full of expensive CRMs that can't do the one thing customers actually want — see what's happening with their deal.

Building from a real problem you've watched up close means you skip the worst trap in software, which is building features nobody wants. I knew exactly what to build because I'd seen the pain.

4. I stopped trying to be perfect

None of these products are "finished." Dinara has bugs I know about. AgentAI doesn't have multi-user team invites yet. MyTradeSite needs a better onboarding flow. They're all live anyway, because users find more bugs than I do, and a live product with a user is worth more than a perfect product in a Git branch.

What this means for you

If you've been thinking about a SaaS idea, or an internal AI tool for your business, or an AI-powered website, or anything else where the bottleneck has historically been "developers cost too much and take too long" — that bottleneck has moved.

It's not gone. Real software still takes real work. But the calculus has shifted enough that things which used to require a £40K agency budget and a six-month timeline can now be built in weeks for a fraction of the price by one person with the right tooling. I'm not the only person doing this. I'm just one of the people who started.

If you're a UK business owner with a problem you've been told would cost too much to solve with software — let's have a conversation. Most of the things you've been quoted £20K+ for I can build in two to four weeks for a fraction of that price. Not because I'm cheap, but because the work itself is genuinely faster now.

If you're another solo builder reading this and wondering whether it's possible to ship at this pace — yes, it is, and you should probably go and try.