Hey buddy,
Two years ago, building a paid web app meant hiring a developer or learning to code.
Now you can describe a small tool in plain English, and an AI builder like Lovable gives you a working app with logins, a database and a landing page. Often in a weekend.
That changes who gets to build software. It does not change who gets to sell it.
Anyone can make an app now. Very few people can find 50 strangers who will pay $19/month for one. This issue is about the second part, with honest numbers on the first.
Btw, my latest project GTMJobsHQ.com is built using Lovable. Read more about it here.
What A Vibe-Coded Micro SaaS Actually Is
A micro SaaS is a small software product that does JUST one job for JUST one type of customer, charged monthly.
Think: a tool that turns client emails into invoices for freelancers, or a dashboard that tracks review scores for restaurants.
"Vibe coding" means you build it by prompting an AI instead of writing the code yourself.
Lovable is one of the biggest tools in the market for this.
You describe the app, it builds it, and you refine it by using and chatting. You can edit and download the code, connect GitHub, and use a custom domain on paid plans.
What Lovable costs:
Free: 5 build credits a day, capped at 30 a month. No custom domain.
Pro: starts at $25/month for 100 credits ($250/year, about $21/month). Custom domains, GitHub sync and the "Edit with Lovable" badge removal are included.
Extra credits on Pro: $15 for 50.
Business: starts at $50/month.
Pro is the one you'd actually use.
For a sense of scale, Lovable's own page lists a sign-up and login flow at about 1.2 credits, and a five-section landing page at about 1.7. A real app takes many prompts, so budget more than one month of credits.
For context, It took me 132 credits for GTMJobsHQ so far (I have topped up twice with $15 worth of credits). This is still far cheaper than any other alternatives out there.
What's real and what isn't:
Founders really are building paying products with AI tools. But the tools are not the hard part.
Here's what I did find:
Launch Fast (Hasaam Bhatti): a non-technical founder who built an Amazon-seller tool and reports about $30K MRR. He used Cursor, not Lovable, and the numbers are self-reported. His first customers came from a partnership with a coaching program. He says himself it's "not a replicable hack for every founder."
Snyho (a solo founder): hit $68 MRR in 3 days from two users at $34/month, then cut the price to $19. Tiny, but real, and he wrote up exactly what went wrong. Again, not Lovable.
Lovable's own case studies (Lumoo, Plinq, QuickTables): they report figures like €700K ARR in nine months. These come from Lovable, which sells the tool. Treat them as marketing, not proof.
So the honest picture: the tools work, a few people do well, and most reported wins are self-reported.
How The Money Actually Works
The model:
You build a narrow tool. You charge a small monthly price. Customers stay as long as the tool saves them time or money.
Illustrative math (my assumptions, not data):
25 customers at $19/month = $475/month
100 customers at $19/month = $1,900/month
100 customers at $39/month = $3,900/month
250 customers at $39/month = $9,750/month
Costs: Lovable Pro ($25/month), domain (about $12/year), email and database add-ons, payment fees, and any AI usage your product itself calls. For a small app that's often $50-150/month, but it grows with usage.
What the reality looks like:
One third-party roundup (unsourced, so treat as a rough guess) says that of micro SaaS products that get any traction, the typical one reaches about $1-3K MRR within 12 months, roughly 10-15% cross $5K, and the top 1-2% pass $10K.
RevenueCat's data, quoted in a separate roundup, says only about 4.6% of new subscription apps reach $10K/month within two years. That's apps, not web tools, but it shows the base rate.
How To Actually Build This
Step 1: Start with the customer, not the app
Launch Fast's founder had tried 10-12 products that never shipped, because he was building for people he didn't belong to. His winner came from a market he actually knew.
Pick a group you understand. Freelancers, Etsy sellers, real estate agents, clinic managers, whoever. Then find a repetitive job they hate.
Step 2: Find a problem people already complain about
Look at Reddit threads, Facebook groups and review sites for the tools they already pay for. Look for lines like "I just use a spreadsheet for this" or "I wish there was a simple tool that..."
If they already use a spreadsheet, you have a product idea.
Step 3: Pre-sell before you build
This is the step most people skip.
Message 10-20 people in that group. Describe the tool in two sentences. Ask if they'd pay $19-39/month for it.
One micro SaaS founder in the research pre-sold 50 lifetime licenses (around $20K) before the product was finished. That founder had 25 years of programming experience, so don't copy the scale, copy the habit.
Step 4: Build the smallest version in Lovable
One core feature. A sign-up. A way to pay. That's it.
Prompt in small steps. Describe the user, the screen, then the action. Fix one thing at a time, because every chat message uses credits.
Step 5: Add payments early
Use a payments tool like Stripe or Lemon Squeezy. Check what your plan and the app's setup support before you promise anything to customers. The integration is very inside and can be made inside lovable itself.
Charge from day one, even if it's $9. Free users tell you very little.
Step 6: Lock down the basics before real users arrive
AI-built apps have a real security problem, and you must take it seriously.
Third-party researchers have reported exposed databases and leaked keys in AI-built apps, including a known flaw (CVE-2025-48757) tied to weak database row-level security. Those numbers come from outside researchers and a vendor blog, and they're not a measure of every app.
What to do: make sure users can only see their own data, never put secret keys in front-end code, and test with two accounts. If you handle payments, health data or anything sensitive, pay a developer for a security review.
Step 7: Get your first 10 customers by hand
Don't wait for SEO.
Show up in the places your customers already are. Share why you built it, not a link drop. Snyho's founder found that telling his story in niche Facebook and LinkedIn groups worked better than posting links.
Offer onboarding calls. Fix what they ask for. Raise the price later.
Step 8: Watch price friction
Snyho cut his Pro price from $34 to $19 after four days because solo users balked. Test one price per audience, and watch who actually buys.
Step 9: Add a referral or partner channel
Launch Fast's biggest source of early customers was a partner with an existing audience. You don't need equity deals, but a newsletter, a community owner or a consultant who already serves your customers can give you your first 50.
Offer a recurring commission, like Snyho's 20% lifetime referral.
Step 10: Plan for what breaks at scale
Once you have paying users, review the code, move to GitHub, and decide whether Lovable is still the right home. One vendor that sells migration services claims feature work gets 2-3x slower as generated code piles up. That's a biased source, but the risk is worth knowing: apps that get messy are harder to change.
Real Numbers Comparison
Launch Fast (Cursor, not Lovable):
Pricing: $49 / $89 / $199 per month
Reported MRR: about $10K at day 30, about $21.8K at day 90, about $30K now
Channel: a partner with thousands of active sellers
Source: founder-written, self-reported
Snyho (not Lovable):
Pricing: $34, cut to $19
Reported MRR: $68 in the first 3 days
Source: founder post
Lovable-built, per Lovable's own case studies:
Lumoo: roughly €700K ARR in nine months (figures vary by roundup)
Plinq, QuickTables: also reported by Lovable
Source: the vendor. Not audited.
Base rate for the rest of us:
Typical traction: $1-3K MRR in a year, if you get traction at all
Reaching $10K+ MRR: the top 1-2%
Realistic target (my estimate): $500-2,000/month by month 12 for a diligent solo builder in a real niche. Higher is possible, but plan for it as an upside.
Common mistakes:
Building before talking to a single customer. The app is the easy part
Choosing a broad idea ("a CRM for everyone") instead of a narrow, painful one
Ignoring security because the app "works" on your screen
Burning credits on vague prompts instead of small, clear instructions
Offering a free plan too soon, then having no way to learn what people will pay for
Relying on one channel or one partner without a backup
Copying the big headline numbers from vendor case studies and assuming they apply to you
Quitting at month three, which is usually before the first real traction
Pro tips:
Pick a niche where you have an unfair advantage: a past job, a community you're in, or an audience you already have
Pre-sell. A paid yes before you build is worth 100 "sounds great" replies
Keep the first version to one feature. If you can't explain it in one sentence, it's too big
Test the app with two separate accounts to make sure nobody sees someone else's data
Export your code to GitHub early so you're never locked in
Track churn from the start. Subscription apps that use AI tend to retain worse, according to RevenueCat's data
Price for the business customer first. $39 from an agency beats $9 from a hobbyist
Reality check:
These are illustrative scenarios, not forecasts.
Month 1-2:
Customer conversations, pre-sales, a first working version.
Revenue: $0-100
Month 3-4:
First paying users, mostly from direct outreach and your own network.
Revenue: $100-500/month
Month 5-8:
Fixing churn, tightening the product, adding one referral or partner channel.
Revenue: $300-1,500/month
Month 9-12:
If it's working, steady growth from word of mouth, content and partners.
Revenue: $500-2,000/month
Year 2 (strong case):
A real niche, strong retention, a few channels working.
Revenue: $2,000-5,000+/month
The other outcome: Many micro SaaS products never get past a handful of users. If that happens, you'll have lost a few hundred dollars and learned how to build and sell software, which is still worth something.
Talk soon, Kris
P.S. - The thing worth remembering about vibe coding.
It made building cheap. It did not make customers cheap.
That means the edge has moved. It's no longer "can you build it?" It's "do you understand a group of people well enough to know what they'll pay for, and can you reach them?"
If you do, a weekend build and $25 in Lovable credits is a very cheap experiment. If you don't, no amount of prompting will fix that.
Start this week. Pick one group you know. Write down three things they complain about. Message ten of them and ask what they use today and what they'd pay for something better.
Worst case: you learn what a real customer problem sounds like.
Best case: you find one that three people will pay $19/month to fix, and now you know what to build.
Quick Message:
As you all know, I do side hustle experiments every month.
I am packaging them into actionable PDF.
The first one is for a cold email agency. it covers everything from how to start, tools to use, niches to target, offers, closing clients, etc.
I think it would be helpful for a lot of you people.
I have priced it at $19/mo, on the last day of every month, you’ll receive such a guide.
The side hustle experiment of this month is about a job board.
Get access to it here:
📫How did you like this idea?
If you’ve read this far, connect with us:
