The problem
Vaultaire launched on the App Store on March 13, 2026, and made $66 in gross sales in its first partial month. “Even before the app was ready, I knew distribution would be the much harder part,” he writes. In the interview he explains that most apps in his category are 15 years old or more and have far more downloads, so he had “zero chance” to outrank them on the App Store alone. His website had its own problem: search engines barely indexed it, and it got 26 clicks in two months.
The blueprint (as he describes it)
- Start with the site and measure everything. A basic website and standard app metadata first, PostHog for tracking, and the domain added to Google Search Console, Bing Webmaster Tools and Yandex Webmaster. He ran a GEO-first skill over the site in loops “until there wasn’t much to do,” which helped “especially around llms.txt.”
- Let the agent read Google’s keyword data. Pulling keywords from Keyword Planner by hand and feeding them to the agent was “a lot of manual work,” so he wrote gum. It lets an agent work with Google services, Search Console and Keyword Planner included. In the interview he says it runs every morning, finds keywords where the site is shown but gets no clicks (his example: shown 16,000 times, no clicks, at about position 50), and writes a guide for it.
- One guide a day, linked from pages that already get traffic. Keyword Planner data and “People Also Ask” questions on Reddit are the seeds. A new guide or scenario goes up daily. He says Google takes three to four days to pick one up, and internal links from pages with traffic point at each new guide. “I literally haven’t written a single line, either of code or the text on the web.”
- Backlinks when nothing gets indexed. He paid two or three directory-submission services a few hundred dollars and republished articles after a delay on WordPress.com, Blogger.com and Medium.com. His domain rating jumped and Google traffic followed, though he says he cannot prove the backlinks alone caused it.
- App Store keywords through an MCP. Astro exposes keywords from Apple’s ad service, and its MCP let the agent work out what to do with them. asccli lets agents change the App Store listing without him opening App Store Connect.
- Translate everything. Each time he translated the app, he translated the website too: every page and every guide, with links swapped for local ones, so a US guide links to the FTC and the EU version to the matching European authority.
- Try other channels and keep what works. Codex turned published guides into visual guides with ChatGPT Image 2 for LinkedIn, X and Pinterest. He also hands Apple Ads to an agent that uses the Astro data. Reddit “went quite poorly,” and he stopped.
Stated results
- $66 in the first partial month (March 2026). In July: 55,000 unique website visitors, 342,000 App Store impressions, 17,100 downloads, 1,287 purchases (a 7.53% download-to-paid rate) and $26,384 in gross App Store sales.
- “Vaultaire.app made $83,530 in October 2026 with a profit margin of 78% (Apple’s cut),” and its first $4k day (X post, Oct 31, 2026).
- “ASO is now responsible for around 60% of all my downloads.”
- Japan “now accounts for nearly half of all App Store traffic and generates thousands of dollars in sales each month.”
- Pinterest: 311 clicks to the site, 244 downloads, 51 paid conversions and $1,122, “All without paid distribution.” LinkedIn and X “produced almost nothing.”
- Monthly costs he lists: Codex $200, Claude $200, Cloudflare $5, Apple Ads about $700.
Caveats
- Every number is his own, from his posts and an interview, and none is audited. The July figures are gross App Store sales.
- gum is his own free tool, so the post also shows off something he built. He doesn’t sell Astro, asccli or Pencil, and he names a friend’s team behind Pencil.
- He doesn’t split the growth by channel beyond the 60% from App Store search, Japan’s share and the Pinterest numbers. He can’t prove the backlinks caused the jump in Google traffic.
- The interview quotes come from an auto-generated transcript of the Starter Story video. The written article is the primary source.
- He is a career engineer who has built half a dozen companies, so the coding agent filled gaps for someone who already knew how to ship software.
- Handles: X @synopsi (bio links vaultaire.app and turek.co; his own article), @vaultaire_app (named in his article), LinkedIn and GitHub ehmo (gum). No Bluesky, Threads, Instagram or TikTok profile verified.
FAQ
How does Rasty Turek use AI agents to grow Vaultaire?
Agents read Google Keyword Planner and Search Console through his tool gum, and a new guide goes up on the site every day. An agent also reads App Store keywords through Astro’s MCP, and agents handle App Store Connect, translations and Apple Ads.
Which channel brings most of his downloads?
App Store search. He says ASO is responsible for around 60% of all downloads.
What didn’t work?
LinkedIn and X “produced almost nothing,” Reddit “went quite poorly,” and his new site got 26 clicks in two months until backlinks from directories and republished articles.
How much does he spend?
He lists Codex at $200, Claude at $200, Cloudflare at $5 and Apple Ads at about $700 a month, plus Apple’s 15% commission.
Sources
- https://x.com/synopsi/status/2084744044392190064 (primary: Rasty Turek’s first-person X article “From $66 to $26,384 a Month: Everything I Learned Building My First iPhone App”, Aug 4, 2026)
- https://x.com/synopsi/status/2116647163564605519 (his X post with the October 2026 figure, Oct 31, 2026)
- https://www.youtube.com/watch?v=NuYO7DWzlaM (Starter Story interview, “I built an app that makes $2K PER DAY”)
- https://github.com/ehmo/gum (gum, his CLI and MCP server for Google APIs)
Want agents like these working your distribution? Get It Built. Get new teardowns weekly: subscribe (double opt-in).
Get one teardown a week.
Same format, a new operator every issue. Double opt-in.