I am not a marketer, and I did not set out to build a company. I set out to solve one small, specific annoyance: small businesses with hundreds of glowing Google reviews and nothing to show for it but a map pin. What started as a weekend idea turned into favURL, and I want to write down, honestly, how it happened, because I think the "how" matters as much as the "what."
The $20 start
I bought the domain for less than $20. No funding, no team, no roadmap deck. Just a name I liked and a problem I could not stop thinking about: every small business I looked at had already written its own best marketing copy, one customer review at a time, and it was sitting there completely unused.
That is the whole entrepreneurial thrill for me. Not the idea of building something big on day one, but the fact that for under the price of a dinner out, anyone can start. I am genuinely glad to be on this journey, small budget and all.
Learning to vibe code
I had never built a product this way before. I leaned entirely into what people now call "vibe coding": working conversationally with AI to design, build, and ship, rather than writing every line by hand from a blank file. It was a real learning curve, not a shortcut. I had to learn how to describe a problem clearly enough for AI to help solve it, how to review what it produced instead of trusting it blindly, and how to keep a growing codebase coherent as I went.
It took more than five weeks to get the first real version live: a working search, a review-reading pipeline, a generated page, and a URL you could actually send someone. Five weeks of long evenings, most of them after a full day of other work. Not fast by startup lore standards, but fast for one person with no team and no code written the traditional way.
The tech stack
For anyone curious what "vibe coded" actually ran on under the hood, here is the real stack:
- Framework: Next.js 15 (App Router, edge runtime)
- Hosting: Cloudflare Pages, deployed via Wrangler
- Database: Supabase (Postgres + Storage)
- Auth: Supabase Auth, magic link and Google OAuth
- AI: Anthropic's Claude API, the actual engine that reads reviews and writes each page
- Business data: Google Places API
- Payments: Stripe, subscriptions and webhooks
- Email: Resend, for welcome emails and page-ready notifications
- Styling: plain CSS with design tokens, no framework, no shortcuts there either
Nothing exotic. The unusual part was not the stack, it was building all of it solo, mostly through conversation with AI, in five weeks.
How it grew from there
The first version did one thing: search a business, read its Google reviews, generate a one-page site. That alone was the hard part to get right, because a generated page that reads like a template is worse than no page at all.
Once that worked, I kept pivoting and layering on what businesses actually asked for. The full feature list, as it stands today:
- AI-generated landing pages from real Google Reviews, structured into a hero, advantages, praised aspects, highlights, special touches, and practical tips, not a generic template
- Photo captioning and galleries, pulled straight from Google Places
- Claiming and ownership, so the business the page is about can take it over via email or Google sign-in
- Full owner editing, inline text edits, drag-and-drop card reordering, section hide and reorder, and single-section AI regeneration
- A real dashboard with sites, analytics, billing, and referrals
- Custom SEO controls per page: title, description, CTA, and dynamic Open Graph images generated on the fly
- JSON-LD structured data so pages show up properly in search
- A directory browsable by category and by city
- A blog and markdown twins of every public page, so AI answer engines can read and cite favURL pages directly
- Paid plans (Basic, Pro, Agency) for businesses and agencies that wanted more pages, no branding, or custom domains
- A referral program, an admin CRM for outreach, and a whole internal Studio tab most users never saw
None of that was planned on day one. Every piece was a response to an actual person asking for it, or an actual gap I found using my own product.
Why I am writing this down
Mostly because I want a record of it, not a polished retrospective years later, but a plain account written while it is still fresh: what it cost, what it took to learn, and what it turned into. If you are thinking about starting something with a domain that costs less than a dinner and an AI tool you have not fully learned to use yet, that is exactly how this one started. I am glad to be on this journey, wherever it goes next.