5 prompts to scale an app from $0 to $10k a month
Find the gap, tear down the competitor, write the PRD, kill the vibecoded look, then solve the revenue equation backwards. Five prompts, each one feeding the next.
Most people start with an idea. Start with other people's money instead. Somebody in your category is already charging for something and already annoying their users about it. That gap is your product, and you can find it in an afternoon.
Five prompts below. Gap, teardown, PRD, design, funnel. Copy each one into Claude, fill the brackets, and don't skip ahead. Every prompt produces a document the next one eats.
Prompt 1: Find the gap
You're looking for apps that make real money from users who are actively irritated. High revenue plus low rating is a market telling you exactly what it wants and not getting it.
Pull the App Store top charts for [CATEGORY], free, paid, and top grossing.
I want the top 100 from each. For every app, get me: price model, rating,
rating count, and how long it's been on the chart.
Then find the mismatches. I'm hunting for apps that rank high on the GROSSING
chart but sit under 4.2 stars, or whose recent reviews keep complaining about
the same three things: price, ads, a missing feature, whatever it is. Those
are apps making real money from users who are actively annoyed.
Give me a ranked table: app, what it charges, the complaint pattern from its
last ~50 reviews, and a one-line gap I could exploit. Rank by
(revenue signal x user frustration). I want 5 targets, not 50.Five targets is the point, because a list of fifty is a list you never act on.
Prompt 2: The teardown
Pick one target off that list. Now read every angry review they've ever gotten. Their one-star reviews are your feature spec, written by people who already pay for this category.
Deep-dive [TARGET APP]. Read its most recent 200 App Store reviews and cluster
the complaints: what do people hate, what do they beg for, what made them cancel.
Then map the paywall: every feature behind it, and exactly what it costs monthly
and yearly.
Now build the undercut plan:
1. Which 20% of their features drive 80% of the value. That's my MVP, nothing else.
2. Which paid feature I should give away FREE, the one that makes their users
feel dumb for paying.
3. My price: aggressive enough to hurt them, high enough that $10k/mo doesn't
need 10,000 users. Show the math at 2-3 price points.
4. A positioning one-liner: "[Their app] but [the thing their reviews beg for]."
One page. If a feature doesn't answer a complaint from their reviews, it doesn't exist.Prompt 3: The PRD, in plan mode
Don't tell Claude to build you an app. Make it write the PRD first, and keep it in plan mode so nothing gets built that you haven't approved. This is the difference between a project you can still maintain in week six and a pile of code you're scared to open.
You're my lead engineer and PM. We're building [APP], [one-liner from Prompt 2].
Use plan mode for this entire project: before ANY code gets written, at every
phase, you present a plan and wait for my approval. Never skip that step, even
when the next move feels obvious.
Step one is a PRD, saved as PRD.md. This becomes the source of truth for
everything after:
- Problem and the one-liner from the teardown
- Target user and the exact complaint we're solving (quote the reviews)
- The ONE core loop a user repeats daily
- Every screen, the data model, and what's behind the paywall
- Non-goals: the features we are explicitly NOT building, so scope can't creep
- Success metrics: what numbers tell us the loop works
Stack: React Native + Expo, Supabase for auth and data. Boring on purpose,
a solo dev has to maintain this.
Once I approve the PRD, plan then build in this order: core loop, onboarding,
paywall. Each phase: plan first, my sign-off, then code, then tests green before
we move on. Anything tempting that's not in the PRD goes in FUTURE.md, not in
the app. Real empty and error states on every screen, and when something fails
the copy names the real cause.The non-goals section is the one people delete. Leave it in. It's the only thing standing between you and a settings screen with fourteen toggles nobody asked for.
Prompt 4: Make it not look vibecoded
This is the step everyone skips, and it's why their app screams that AI made it. A design system before any screen gets touched, then the three moments that actually sell: first open, the core loop, and the paywall.
You're a senior product designer doing a full pass on my app. First, build the
design system. Don't touch a screen until it exists: a real palette (near-black
base, ONE accent color, semantic colors used only for meaning), one type family
with a proper scale, spacing on a 4pt grid, one corner radius language. Write
them down as tokens.
Banned: default blue, system-font energy, gradients on everything,
purple-to-pink hero sections, emoji as icons, five different button styles.
Then redesign every screen against those tokens, and obsess over the three
moments that sell the app: first open, the core loop, and the paywall. The
paywall should feel like a VIP invite, not a ransom note, with price and renewal
terms stated plainly. That honesty converts AND keeps you out of trouble.
Deliver clickable HTML mockups of the 5 core screens so I can walk through them
before we touch the real code.Mockups before code, every time. Redesigning an HTML file costs you ten minutes. Redesigning a built app costs you a weekend and half your motivation.
Prompt 5: The $10k math and the weekly loop
Ten grand a month is just an equation. Make Claude solve it backwards so you know exactly which number you're attacking on any given Monday.
My app charges [$X/mo, $Y/yr]. Work the equation backwards from $10,000/mo:
how many paying subscribers is that, and, using realistic freemium conversion
rates rather than fantasy ones, how many installs does that require, and how
much traffic does THAT require. Show the full funnel with the numbers at each
step, then tell me which single conversion step is most worth attacking first.
Then build the growth system:
1. ASO: 30 keyword targets ranked by (volume / competition), plus title and
subtitle variants to test.
2. Screenshot copy: 5 captions that sell the outcome, not the features.
3. 10 short-form hooks for TikTok and Reels: build-in-public angle,
before/after angle, and "I found a gap" angle.
4. The weekly loop, as an actual checklist. Every Monday: read the week's new
reviews, ship the #1 complaint, post 3 videos, check the funnel number we're
attacking. Repeat until the equation balances.That last checklist is the whole business. Read the reviews, ship the top complaint, post the videos, watch the one number.
Why the order matters
The prompts aren't the magic. The magic is that each one produces a document the next one runs on. Gap into teardown, teardown into PRD, PRD into design, design into funnel. Run them out of order and you get five disconnected chat sessions. Run them in order and you get an operating system.
