Notes
Building solo, augmented by AI
I wrote code for fifteen years, from front-end integrator to React lead developer. Since then, I have designed and built nine SaaS products on my own, with AI as my extra hands: an ATS with AI scoring, a semantic sourcing engine, an agent that detects sales opportunities, a fully local meeting recorder, a gallery and payment platform for photographers…
These notes are what that work teaches me. What an MVP really costs when a senior developer leans on AI. What AI speeds up, and what it does not replace: framing, architecture, the judgement about what not to build. How to bring it into a team without creating debt. And, on the acquisition side, how a landing page turns an offer into booked calls.
No growth promises, no made-up numbers. Just what I built, what it cost in time and decisions, and what I would do differently.
All notes
How much does a SaaS MVP cost in 2026, with a senior developer augmented by AI?
Sourced market ranges, and what actually drives the price: scope, payments, multi-tenancy, AI. What AI changes in the budget, and what it does not.
· 5 min read
How I build a SaaS solo with Claude Code: method and guardrails
Frame before coding, a CLAUDE.md that tells the truth, rules the agent cannot bypass, and what I never delegate. The method behind nine products.
· 3 min read
Bringing Claude Code into a dev team without creating debt
The debt a coding agent creates, and the guardrails that prevent it: written context, executable rules, specs before prompts, docs treated as code, humans in the right places.
· 3 min read
A landing page for tradespeople: turning a workshop trade into a quote machine
A promise about the customer, the quote on the first screen, fears answered before they are voiced, honest proof and useful local pages. The Autovring case, broken down.
· 3 min read
A product to build?
Framing, architecture and development, augmented by AI, at a fixed price per deliverable milestone. Let’s talk for 30 minutes.
See the AI product build offer