Most side projects make exactly $0.
You build the app, you tweet the launch, your mom likes the post, and that's the whole revenue story.
So when the Starter Story research team sat down and pulled apart three side projects that make money right now ($16K, $70K and $15K a month, all started by one person with a day job), I paid attention. Pat Walls runs this exercise every week with his researchers Gus and Marco: find what's actually working, run the numbers on camera, no guru pitch.
Watch the full episode here, then stick with me for the breakdown and what I'd steal from each one:
1. Parakeet AI: $16K a Month From a Weekend Build
Parakeet AI is a real-time interview assistant. It listens to your job interview, transcribes the question, and feeds you an answer while you're on the call.
Is it controversial? Very. Recruiters show up in the TikTok comments saying "we can tell." People argue about whether it's cheating. The founder kept posting anyway.
That founder is Jure Sotosek, a developer from Ljubljana, Slovenia. He built the MVP in a weekend in September 2024: transcribe the call, send it to ChatGPT, collect payment through Stripe. He paid an agency about €1,000 for a Webflow landing page and shipped.
First month: $700.
By month three he was at $16K a month and hitting his first €1,000 revenue day.
The growth channel was organic short-form video. One TikTok, filmed at his desk with zero production, pulled almost 3 million views and 156,000 likes. And when you scroll his account, you notice something: it's the same video over and over. Same desk, same laptop, same fake interview. Only the text on screen changes.
Pat ran the napkin math on that one video live in the episode. Say 10% of 3 million viewers visit the site: 300,000 people. Say 1% of those buy the cheapest pack at $30 (three interview credits, no subscription, 30-day money-back guarantee). That's 3,000 buyers and $90,000. Cut every assumption in half and it still made around $45,000.
From one video he filmed in his room in under an hour.
The pricing choice matters too. A job interview is not a recurring problem, so Jure sells credits instead of subscriptions. No churn dashboard to stare at, no cancellation flow to build.
And he's still going: ParakeetAI 2.0 launched on Product Hunt in October 2026 with a full interview-prep suite around the live assistant.
→ The takeaway: find one content format that converts, then repeat it shamelessly until it stops working. Most of his videos did 200 to 400 views. He kept posting until one did 3 million.
2. ProjectionLab: From $200 on Product Hunt to $1M ARR
Kyle Nolan had a corporate engineering job, financial anxiety, and a problem: every long-term financial planning tool he tried looked like it was built in 2006.
His own words, from Indie Hackers: "I couldn't find a long-term financial planning app that felt modern and actually fun to use. I spent a year and a half of my free time building one for myself."
That became ProjectionLab, a planning tool where you plug in your numbers and watch your financial future render as one clean chart. Retirement date, savings rate, what happens if you quit your job next year. The FIRE crowd loves it.
The timeline is the least sexy growth story you'll ever read, and that's exactly why I trust it:
- May 2021: first revenue, $150 MRR, from a Hacker News post
- Dec 2021: $1K MRR (nine months of nights and weekends to get there)
- Sep 2022: $4K MRR after a Mr. Money Mustache shoutout
- Jun 2023: $10K MRR
- Sep 2023: $16.6K MRR, drops his day job to part-time
- Nov 2023: $23.3K MRR, quits entirely
- Jun 2025: $83.3K MRR. One million dollars in ARR, zero funding
At the time this episode was recorded, Gus pegged him around $70K MRR with roughly 100K visitors a month. The average price works out to about $9 a month, which means roughly 8,000 paying customers. There's a free tier, paid plans on top, and an $800 lifetime deal for people who hate subscriptions.
Notice what he never did: no paid ads, no sales team. He built in public on X, posted milestones, and let Hacker News, Product Hunt and the FIRE blogs do the distribution. Today ProjectionLab serves over 100,000 households with a team of three.
Gus made a point in the episode that I keep thinking about: a $9 consumer price point means nothing in the business is ever urgent. No enterprise customer calling you at 9am. No sales calls on your lunch break. That's what makes it survivable next to a full-time job.
→ The takeaway: solve a problem you have, charge a price that doesn't require support, and accept that year one might pay you $1K a month. Kyle worked 4 to 6 hours a night for two years before this was his job.
3. HabitKit: $15K a Month From an Idea That Usually Makes $0
Habit trackers are where side projects go to die.
Pat put it bluntly in the episode: first-time founders launch habit trackers constantly and they make nothing. Even Marc Lou, who has one of the biggest indie hacker audiences on X, has a habit tracker (Habits Garden) sitting at $174 a month.
HabitKit makes $15,000 a month. Same category. Smaller audience. So what's different?
Two things.
First, the visual. Sebastian Röhl, a German developer, took the GitHub contribution graph (that grid of green squares every coder obsesses over) and turned it into a habit tracker. Your meditation streak becomes a wall of colored tiles. Miss a day and there's a hole in the grid. It's the same psychology that keeps developers committing code at 11pm, pointed at your gym habit.
That grid is also a growth engine. A screenshot of your 100-day streak compresses months of hard work into one image, and that's exactly the kind of thing people post. Pat's own version of this: his GitHub contribution graph tweet ("shout out to all the people who said I got lucky") did 2.5 million views.
Second, and this is the part most people skip: it's a mobile app, not a web app. The App Store has hundreds of millions of users searching for things every day. Sensor Tower data pulled up in the episode showed HabitKit doing around 40,000 downloads a month. A web-based habit tracker gets none of that for free.
Sebastian's App Store playbook is deliberate. "Habit tracker" is in the app name itself, he ranks top 5 for the main search terms, he runs about $100 a month in Apple Search Ads partly as a ranking signal, and the app asks for a review right after you complete your first habit. Over 800 reviews and counting. I broke the whole thing down in my HabitKit ASO deep dive if you want the full playbook.
The revenue splits roughly half subscriptions (around $7.8K MRR) and half lifetime purchases (around $7.3K a month). Cheap product, volume play.
And Sebastian is not an overnight story either. He quit his job in April 2022, launched a workout tracker called LiftBear first, watched it stall at $150 MRR, then shipped HabitKit in November 2022.
→ The takeaway: a dead category isn't dead if you change the distribution. The idea "habit tracker" made Marc Lou $174 a month and Sebastian $15K a month. The difference was App Store search plus a shareable visual, not the idea.
What These Three Side Projects Have in Common
Pat closed the episode looking for the common thread, and it maps to what I see on the podcast every week.
They're all consumer products. Not B2B SaaS. Gus's reasoning: a side project can't afford urgency. B2B means sales calls, support SLAs, and being on call. A $9 consumer app means nobody is ever angry enough to need you at 9am.
They all sit in health, wealth or work. HabitKit is self-improvement, ProjectionLab is money, Parakeet is getting a job. Painkillers, not vitamins. Marco's contrast was meal-prep apps: "help me pick a recipe" is a vitamin, "I'm not where I want to be financially" is a pain people pay to kill.
They all started as the founder solving their own problem. Kyle wanted a better planning tool. Sebastian wanted a prettier habit tracker. Jure saw a gap in the job market and moved in a weekend.
And they all repeat one motion instead of chasing new tactics. Jure reposts one TikTok format. Kyle posts milestones on X. Sebastian grinds App Store rankings. None of this is creative. All of it compounds. Once you find the motion that works, the job becomes turning it into a system that runs without you thinking about it (this primer on business automations is a good place to start if you're at that stage).
If you want more ideas in this lane, I keep a running list of micro SaaS examples with real revenue numbers.
FAQ
What kind of side projects actually make money?
Consumer apps in health, wealth or careers, priced low enough that customers don't need support. The three in this episode (an AI interview assistant at $16K/mo, a financial planning tool at $83.3K MRR, and a habit tracker at $15K/mo) all fit that pattern. B2B can pay more per customer, but it demands sales calls and support hours a side project doesn't have.
How long does it take a side project to make real money?
Wider range than anyone admits. Parakeet AI went from $700 to $16K a month in about three months because short-form video can spike overnight. ProjectionLab took nine months to reach $1K MRR and two and a half years before the founder went part-time. Plan for the slow version and treat the fast one as luck.
Can I build a profitable side project while working full time?
Kyle Nolan did exactly that: 4 to 6 hours a night plus weekends for two years, and he only quit his day job at $23.3K MRR. The trick is picking a product with no urgency built in. Low prices, self-serve onboarding, no sales calls. If your side project needs you during business hours, it needs your job gone first, and that's backwards.
Do I need to be a developer to copy these playbooks?
It helps, but less every month. Jure coded Parakeet's MVP in a weekend, but Marco's point in the episode was that current AI tools let non-technical founders spin up an MVP almost as fast. The harder skill is the marketing motion: making two TikToks a day, or posting your numbers in public, or grinding App Store keywords for months.
One more thing. Every week on the Profitable Founder Podcast I interview bootstrapped founders about exactly this: the real numbers, the slow years, what actually moved revenue. If these three stories were useful, the long-form versions are better.
