AI Side Hustles Half-true — works only if you do the unspoken work

By the Income Reality Check Editorial Team · Updated 2026-09-30 · 7 min read

Stella’s $360K/mo manifesting app: what the Claude build tutorial skips

Verdict: Half-true — works only if you do the unspoken work. The app and the revenue are real; the audience that produced the revenue is not something a four-prompt tutorial can hand you.

A video from the channel DIGITAL INCOME PROJECT, with nearly 55,000 views, opens with a striking claim: a woman named Sarah scaled a digital product “from zero to 360,000 a month” in two months, and the host, Clay, promises to reverse-engineer the whole thing and rebuild it live using Claude, Higgsfield, and Zapier. The founder is real, the app is real, and the revenue checks out. What the video soft-pedals is the one ingredient that actually produced the money — and it’s the ingredient a tutorial can’t give you.

What the video actually claims

The app is called Stella. You tell it your goals, and it generates custom visualizations narrated by “your future self who already achieved them,” plus daily affirmations. That’s the whole product — no feed, no twelve features. The founder, Sarah Pearl, told the podcast Starter Story that within two months of launch she hit roughly $300K MRR (climbing toward $360K), 12,000 paying customers, and over 200,000 downloads, with an MVP built in two days on a stack of Claude Code and Google AI Studio (Starter Story).

Clay’s pitch is that you can copy this. He anticipates the obvious objection — “she only got results because of her audience” — and answers it by showing an Instagram account he grew from zero to 10,000 followers and 500,000 views in 30 days using an AI influencer (an older woman doing horoscope and manifestation readings), built with Higgsfield’s image and video tools. Then he walks through building a web-app version of Stella in “less than 10 minutes” with Claude, Zapier, and Netlify, from “just four prompts.”

Here’s the part worth quoting fairly. Near the end, Clay admits he hasn’t priced his app yet and hasn’t collected a dollar. “I’d actually like to test this product out with my audience first… The monetization part is what I’m testing now.” So the video demonstrates the audience step and the build step. The money step? Still a promise.

What the method actually requires

Let’s separate what’s genuinely easy from what isn’t.

Building the app is the easy part, and the video is honest about that. AI coding tools really have collapsed the effort of shipping a simple web form that captures emails and fires daily messages through Zapier. Sarah confirms she used Claude Code; Clay confirms the same. No dispute there.

The revenue is the hard part, and it doesn’t come from the app. It comes from attention. Sarah didn’t launch to strangers — she launched to an audience of roughly 1.2 million Instagram followers and 2.8 million on TikTok, with over a billion cumulative views (Starter Story). Her own framing, quoted in the video, is telling: “If you have a personal brand and if you have a following… I could sell air and it’ll sell.” That’s not a footnote to the strategy. That is the strategy.

Now weigh Clay’s proof against the promise. He grew an account to 10,000 followers and half a million views — genuinely fast, and a fair demonstration of the AI-content workflow. But 10,000 is not 4,000,000. And he did it with a distribution trick he mentions almost in passing: “I had other accounts in our niche re-sharing every post.” Owning or partnering with a network of pages that amplify your content is itself an asset most beginners don’t have. It also took him 17 uploads before a single video cleared 2,000 views. The overnight framing hides a grind.

Then there’s the economics of subscription apps, which the video never touches. RevenueCat’s State of Subscription Apps 2026, drawn from tens of thousands of live apps, found the cross-category median revenue one year after launch is about $72 a month. Only 17.3% of apps ever reach $1,000/month, and just 4.6% hit $10,000/month within two years (RevenueCat). AI apps have a specific weakness on top of that: they earn 41% more per payer but churn about 30% faster. Manifestation is a mood purchase. Keeping 12,000 people paying month after month is a retention problem the tutorial doesn’t mention.

Who actually wins this game?

The winners here are people who already own distribution, or who have the rare, real skill of manufacturing it. Sarah says as much — she reframes the question from “do you have an audience?” to “do you have the skill of virality?” That skill is genuine and valuable, but it’s not a four-step checklist. CNBC has reported that creators are flocking to subscription apps precisely because content income is volatile, and the ones who succeed are those who can point an existing, engaged following at a paid product (CNBC).

The creator economy is brutally top-heavy. Analyses of paid partnerships and platform payouts consistently show a winner-take-most curve — a thin top tier captures the bulk of the money while the majority earn little, a pattern the writer Peter Yang has documented across Twitch, Substack, and Spotify (Behind the Craft). Sarah is at the top of that curve. Most people running the exact same playbook will be somewhere in the long, quiet bottom.

What you’d realistically earn

If you start from zero with no audience and no amplification network, the honest range for the first six months is close to nothing — think $0 to a few hundred dollars a month, matching the RevenueCat median rather than the headline. Cross into meaningful money, and you’re in the 17% of apps that clear $1,000/month; that typically follows a year or more of consistent content and a following in the tens of thousands, not the ability to write four prompts.

One structural note the video gets right, and it’s a real edge: shipping a web app instead of an iPhone app sidesteps Apple’s App Store commission — 30%, or 15% for developers under $1 million through the Small Business Program — and skips the review queue entirely. That’s a legitimate margin advantage. It just doesn’t create demand; it only protects the demand you’ve already earned.

Who this is (and isn’t) for

This makes sense if you already post consistently, understand what makes a reel get saved and shared, and have 10–20 hours a week to feed the content machine for months before expecting revenue. If you have a small but real following in a high-emotion niche — manifestation, fitness, dating, money — bolting a simple paid app onto it is a sound move, and the build tools are as good as advertised. It does not make sense if you’re picturing a passive, set-and-forget income from four prompts. There is no product-only version of this business. The product is the easy 5%.

One more thing worth knowing if you scale and start advertising specific earnings: the FTC has proposed a new Earnings Claim Rule that would require written substantiation for money-making claims and let regulators seek refunds and civil penalties for deceptive ones (FTC). U.S. readers who sell a “copy my $360K app” course, in particular, should keep receipts.

What to remember

Stella is a real app doing real money, and Claude really did build it — that much is not hype. But the $360,000 was purchased with four million followers and years of practiced virality, and the video’s own host hasn’t yet turned his 10,000-follower rebuild into a single dollar. Treat this as an accurate demo of the build and an unfinished experiment on the money. For more on where AI products actually pay, see our looks at simple AI digital products people are selling and Claude side hustles that can pay a full-time income.