Everyone keeps telling me SEO is dead.
Then I watch a founder like Rasty Turek pull up his App Store dashboard. His encrypted storage app made $66 in its first month. Six months later it did $41,000 in a single month. It now averages $2,000 to $3,000 a day.
No TikTok, no paid ads, no UGC creators.
One AI agent, publishing one SEO guide per day, while Rasty sleeps.
He broke the whole thing down with Pat Walls on Starter Story, and it's the clearest picture I've seen of what app marketing looks like when you hand the boring parts to a coding agent. Watch it here:
A Dropbox even the FBI can't open
The app is called Vaultaire. Rasty describes it as Dropbox with one caveat: everything is encrypted on your device. You draw a pattern on a grid, and that pattern becomes the encryption key for your vault. Photos of IDs, birth certificates, crypto wallet keys, anything you'd rather not leave sitting in your camera roll.
No account, no servers holding your files, no shared keys. Rasty can't see your data, and neither can Apple.
He found that out the hard way. One Friday he got a call from an FBI agent. They'd caught someone they believed had used the app to store sensitive information, and they wanted Rasty's help opening the vault. He couldn't. The app is built so he can't even tell whether a vault has anything in it.
The agent was not thrilled. But that phone call is the product working exactly as designed.
And here's the part that matters for you: this is not a novel idea. Encrypted vault apps have existed for 15+ years. Rasty picked a proven category where people already search for the thing and already pay for the thing. Same logic I covered in boring SaaS: you don't need a new idea, you need a better wedge into an old one.
From $66 to $41,000 a month
Rasty's backstory is wild. He grew up in a Slovak village of 700 people. He says he moved to Boston for a security company that got acquired by Google, worked on the Gmail team, then left to start his own companies. One got acquired by Amazon. His last one, Pex, was a venture-funded copyright platform for the music industry that he ran for over a decade before selling it in 2025.
So yes, he's a real engineer. But Vaultaire is his first iOS app ever, and he didn't write it by hand.
The numbers he shared on camera:
- Built in roughly 3 weeks with coding agents (Claude and Codex)
- Launched in March. First month: $66
- Month six: $41,000
- Now averaging $2,000 to $3,000 a day
- 60,000+ total downloads
- Pricing: $1.99 a week, $19.99 a year, $39.99 lifetime
The pricing split surprised me. About 60% of his purchases are lifetime plans, and they account for north of 80% of revenue. He almost shipped the app completely free. The day before launch he slapped a price on it, mostly so support requests wouldn't eat his life.
That last-minute price tag is now the entire business.
What agentic SEO actually looks like
Here's the quote that stopped me: "Everything is done by one of the coding agents."
Rasty hasn't written the code, and he hasn't written the marketing site either. Not the feature pages, not the guides, not a single line.
His reasoning was sharp. The incumbent vault apps are 15 years old with millions of downloads, so Apple ranks them first in the App Store and he'd have zero chance there. But those same incumbents ignore the open web. They're big enough that they don't have to care.
So he went where they weren't: he decided to own every search query that points at his category, on Google instead of the App Store.
The system behind it, step by step:
→ He wrote an open source command line library called gum: MCP tools on top of Google services, including Search Console data and Keyword Planner.
→ Every morning, Claude wakes up and runs the pipeline. It looks at yesterday's search data and finds keywords where the site got impressions but no clicks. His example: your site was shown 16,000 times for a keyword, but you're sitting at position 50, so nobody clicks.
→ The agent writes one guide targeting that keyword. Things like "how do I store my photos securely on iPhone." One per day, no more.
→ It wires internal links from existing high-traffic pages into the new guide, which is how Google starts taking the new page seriously. Indexing takes 3 to 4 days.
→ Every guide has a CTA placed where readers actually stop scrolling, pushing them to the App Store.
The result: the web is now the majority of his traffic, and it converts to the app at around 10%.
Total marketing team: zero humans.
Why one article a day beats a thousand
This is where most people get agentic SEO wrong.
The obvious move with an AI content pipeline is volume. Spin up the agent, publish 1,000 pages overnight, pray. That's the slop strategy, and Google has been burning those sites down for two years.
Rasty does the opposite. One guide a day, each one anchored to a real keyword with real impression data behind it. In his words, he makes sure the content is "not just like vomit of words." Google seems to reward the cadence: a steady drip of pages that each earn their spot.
Gus, Pat's co-host, pushed back on the whole premise at the end of the episode: "I still think SEO is dead, whatever you want to call that." His take is that it works for Rasty because the niche is so specific, and because a privacy app earns word of mouth and LLM recommendations that Rasty can't even track (he refuses to put heavy analytics in a privacy product).
Honestly, both things are true. Rasty's guides feed Google today, and the same authority feeds ChatGPT and Gemini recommendations tomorrow. Niche sites with genuinely useful pages are exactly what LLMs cite. I've seen the same pattern in the Reddit and SEO playbook Elephas used: unsexy channels, compounding quietly, while everyone else fights over TikTok.
The conversion curve nobody optimizes for
One more number worth stealing.
Vaultaire converts 5.47% of downloads to paid on day one. Most trendy apps (Rasty name-checked Cal AI) are built to convert you instantly or lose you forever: hard paywall, countdown timer, now or never.
His app has the lowest conversion on day one. It peaks around day 60.
Think about why. You download an encryption app, you test it with a few photos, and you slowly move your actual life into it. The more you store, the more a $39.99 lifetime plan feels like nothing. Trust takes weeks, and his pricing is patient enough to wait for it.
If your product genuinely gets more valuable with time, copying the hard-paywall playbook from consumer AI apps might be the most expensive mistake you can make.
Steal this playbook
If I were launching a utility app this month, here's the Rasty version of the plan:
→ Pick a proven, boring category where the incumbents are old and lazy. Encryption, converters, scanners, trackers. Demand already exists.
→ Build it with coding agents in weeks, not months. Vaultaire took 3 weeks. Ship, then improve.
→ Skip the App Store knife fight. Build a marketing site and go after long-tail search queries the big players ignore.
→ Set up an agent loop on your search data: find impressions with no clicks, publish one quality guide a day, internal-link it from pages that already rank.
→ Price for trust. Offer lifetime. Let the conversion curve run past day 30 before you judge it.
→ Then do the hardest part, which Rasty says is just starting: "as long as they're not out there, they don't exist." He has two dozen project ideas in his head. The one he shipped pays him $2,000 a day.
FAQ
What is agentic SEO?
Running your SEO operation through AI agents instead of a content team. In Rasty Turek's setup, a coding agent reads his Google search and Keyword Planner data every morning, finds keywords where his site gets impressions but no clicks, writes one guide targeting that gap, and publishes it with internal links. No human writes or edits the content.
How much money does Vaultaire make?
Per the numbers Rasty shared on Starter Story: $66 in its first month (March), $41,000 in month six, now averaging $2,000 to $3,000 a day with 60,000+ downloads. Roughly 60% of purchases are $39.99 lifetime plans, which drive over 80% of revenue.
What tools does Rasty Turek use?
The app is native Swift with UIKit on top of iCloud. He builds with Claude and Codex, runs his SEO loop through gum (his open source CLI exposing Google services to agents via MCP), and uses PostHog's free tier for measurement. The stack costs close to nothing.
Does SEO still work for apps in 2026?
For niche utility apps, yes, and it's getting a second life. Rasty's web traffic converts at around 10% and outgrew his App Store traffic. The same authority that ranks you on Google also gets your app recommended by ChatGPT, Gemini, and AI search overviews, which is where a growing share of installs starts.
Want more stories like this?
Every week on the Profitable Founder Podcast I sit down with bootstrapped founders and get the real numbers: the $66 first months, the accidental paywalls, and what finally worked.
Listen to the latest episode