The Cottage Gig
My sister-in-law and her husband have a cottage in Guthrie, Oklahoma. Their current website was built by well-meaning folks doing their best with small-town talent—I mean no disparagement. But looking at it, I saw an opportunity to flex Claude Code’s muscles.
The original site: pembrokecottageguthrie.com
What I built in 2 hours: plentyofsaas.github.io/pembrokecottage
Two hours. While watching Squid Games Challenge. That’s the reality now.
Here’s what’s interesting: the original site may have cost up to $1,000, and they’re paying ongoing hosting fees plus a maintenance retainer. My build cost a few tokens and lives on my $4/month GitHub Pro plan.
Now granted—I didn’t spend time on discovery, consultation, or developing site architecture. But property rental sites have a canonical form. I asked Claude Code to adopt the persona of an expert property listing designer for Airbnb or Vrbo. The photos aren’t great (I didn’t touch them), but I improved the viewing experience by adding a lightbox.
The Business Thought
As I sit here between paychecks—or maybe having been paid my last one—I think about opportunities.
How many mostly-static sites out there need a refresh and cheap hosting? Dentists. Cottage rentals. Local restaurants. Small professional services. The long tail of businesses paying $100+/month for something that could cost $4.
It’s not a business that will fulfill me. But if I tested into some demand, found some, and got a few up and running? It extends the runway. Just thinking out loud.
The Landing Page Stack
I’m also exploring a problem to solve. The approach: a landing page that lets folks opt into an “early access” group via a form.
By “early access,” I mean these folks will get concierge fulfillment—I do the work behind the scenes to inform how (and whether) to build the product. It was before my time, but they say the original Stitch Fix customers were being styled with spreadsheets. In my Geoforce days, the entire company helped pack hardware in afternoon packing parties.
You do what you gotta do to get to the next step toward scaling. It’s very often not sexy.
What did I build this landing page in?
- Jekyll - Static site generation
- GitHub Pages - Free hosting
- Cloudflare Workers - Serverless form handling
- Postmark - Transactional email
- PostHog - Analytics
This is my $20/month stack where I have several experiments running. And I did a lot of this work sitting in the lobby of an automobile shop, getting my oil changed.
The Tip
As a backend engineer with frontend taste but slightly-below-serviceable design skills, Claude Code’s frontend capabilities are a game-changer.
If you want to level up, read this: Improving Frontend Design Through Skills.
But here’s the caveat: you need to have product taste. AI can execute design, but it can’t tell you what should be designed. If you have taste—if you know good when you see it—it’s now game on for you.
The gap between “I know what I want” and “I can build what I want” just collapsed.
What’s Next
- Keep testing the cottage site approach—might reach out to a few local businesses
- Monitor the landing page for early access signups
- Continue building in public with whatever runway I have
No guarantees on any of this. Just experiments. Just extending the runway one day at a time.
Built with: Claude Code, Jekyll, GitHub Pages, Cloudflare Workers, Postmark, PostHog Cost: ~$20/month for the entire experimental stack Lesson: The gap between taste and execution has collapsed—if you have taste