The vibe-coded case at a glance
- What he sells: ChatSEO, positioned as "ChatGPT but for finding hidden SEO opportunities"
- Idea to paid beta: 3 months, non-technical founder, no-code AI stack
- Weekend validation: 200-300 emails, 35% landing-page click-to-signup
- MVP build time: 2 weeks
- Recorded beta discovery calls: 25, via Fathom, piped into NotebookLM
- MRR after paywall drop: $500 in 3 days on a 550-person beta cohort
- Public launch: Product Hunt, Show HN, Tiny Launch, Uneed all still queued
Stack: V0, Lovable, Base44, Replit, Vercel, Figma, Fathom, NotebookLM, Stripe. Organic distribution on LinkedIn, Reddit, and X. Cost basis rounds to a domain and a Vercel tier. He is his own ICP: an SEO practitioner who wanted an AI-native tool and couldn't find one he liked, so he shipped the one he wanted.
What the AI SEO tool is actually selling
The product isn't the chat interface. The product isn't even the SEO workflow. The product is a 25-conversation, weekly-growing customer-voice corpus that nobody else in his category owns.
Here's the mechanism. He recorded every beta discovery call with Fathom. He dumps the transcripts into NotebookLM. Every landing page rewrite, feature spec, cold DM, ad hook, and objection-handler now comes out of a queryable archive phrased in his ICP's own words. His copy hits because it's not copy. It's his customers' language reflected back at them at 30% click-to-signup rates.
Eighteen months from now there will be five vibe-coded ChatSEO clones. The MVP is copyable. The three-wave beta with a paywall dropped at wave two is copyable. The 7-channel sequenced launch is copyable. The corpus is not.
He thinks he built a productivity trick. He built the moat.