How Vaultaire grows

He wrote GUM, open-source command-line tools with an MCP layer over Google services including Google Search and Google Keyword Planner. Every morning Claude (or Codex) runs through the data, finds keywords where the site is shown but gets no clicks (e.g. 16,000 impressions at position ~50), writes an optimized guide for that keyword and publishes it. Cadence is about one guide a day, deliberately not bulk AI content; he checks that guides are high quality, not 'vomit of words'. Google takes 3-4 days to register a new guide. He says 99% of the remaining work is this morning run.

He planned to release the app completely free, then the day before launch decided to charge a few dollars for some advanced features, mainly so paying users would justify the support requests. Plans: weekly $1.99, annual $19.99, lifetime $39.99. The core app stays usable for free, and unlike apps that push to convert on day one, users convert over weeks.

The category's apps are 15+ years old with millions of downloads, so Apple ranks them first and a new app has no chance in App Store search, but they neglect the web. He set out to rank on the web for every keyword that points to the same thing as the App Store searches: first a basic marketing site outlining all features, then a series of how-to guides such as 'how do I store my photos on iPhone'. All code and copy are written by coding agents. Authority built over time, bringing traffic that converts to downloads. No paid ads and no paid creators.

After each new guide is published he optimizes internal links so pages that already get traffic point to the new guide, which he says is how Google comes to prioritize it.

Once traffic came, he focused on conversion: he tells the coding agents to structure each article's CTA so that wherever readers commonly drop off in the article, they can go straight to the app.