Can you still make money indie hacking?

Yes. People still make money from products they own. The honest second sentence is the one the feed skips: most products that show up in public make almost nothing, and the ones that look like a salary are a thin slice, often reported by the founder.

Indie hacking is that bet. You own the product. Someone pays. A venture round is not what makes it count. The question “can you still” usually means “does the 2019–2023 shortcut still print money?” The answer to that narrower question is no. The answer to the living is yes, with ugly odds.

What “make money” is being asked to mean

A first sale is money. A thousand dollars in a good month is money. A salary you can leave a job for is a different claim. The timeline treats all three as the same screenshot.

Monthly recurring revenue is the number the culture posts because a subscription dashboard makes months comparable. It is not profit. It is not a year. It is not what a person takes home after tax, tools, refunds, and the hours. A product at $1,000 MRR is about $12,000 a year if the month holds. In most places that is a proof, not a living. The first $1,000 MRR page is about that proof. It is not a finish line.

Profit is the quieter number. On 7 September 2026 Marc Lou said, on the landing page for his free book, that a founder-reported $3 million across about ten years and 36 startups came with margins around 85%. That describes a small shop that keeps most of what it earns. It does not say which of the 36 products earned it. The story of that pile is the useful cut: a decade, not a launch week.

Pieter Levels has put operating revenue in a band around $200,000 to $250,000 a month, and he has claimed a profit margin around 94.5% after cancelling a stack of SaaS bills. The $10 million line he posted on 23 September 2026 is revenue plus investment gains, not the shops alone. Read the mix. Do not flatten it into “an indie hacker makes $10 million.”

What the datasets actually show

None of these is a census of indie hackers. They are the public numbers that sit next to the question, with the limit attached.

Carta’s Solo Founders Report 2025 says the share of new U.S. startups on Carta with a solo founder rose from 23.7% in 2019 to 36.3% in the first half of 2025. That is cap tables, not a directory of micro-SaaS. More people are incorporating alone as starting gets cheaper. Incorporation is not revenue.

Stripe Atlas says 20% of Atlas companies now reach a first paying customer within 30 days of incorporation, up from 8% in 2020. First revenue got faster. A first customer is not a living.

ChartMogul’s 2025 growth report finds newer SaaS cohorts slightly worse at reaching $1 million ARR within three years than the 2016–2019 group. The 2021 cohort is the weak one, at 10.1%. Easier tooling did not buy better odds of scale. ChartMogul is subscription analytics for companies that connect it. It is not a vibe-coding census.

At the indie-visible end, the shape is a pyramid. Steal What Works compiled a TrustMRR snapshot in which 67.8% of businesses sat under $1,000 in lifetime revenue and 0.9% sat over $1 million. Lifetime revenue, not MRR. A point-in-time pull. Founders choose to connect Stripe. Young projects are over-represented. The live board moves. The direction is still the one to remember: almost everyone is at the bottom, and the screenshot culture photographs the top.

Primary sources for Carta, Stripe Atlas, and ChartMogul. The TrustMRR split is a compilation of a moving verified-revenue feed. The longer table lives on revenue benchmarks.

Why the feed says the money stopped

X shows people who post. People post when a number is large enough to stop a thumb, or when a small number is part of a story they are selling. Quiet months do not travel. A product that sells through a niche inbox, a search page, or a community the founder already had often has nothing to screenshot, so it drops out of the argument about whether money is “still” possible.

The old loop had slack in it. Code was slow. A Product Hunt day could move a launch week. Informational pages could send clicks for years. Pew watched real Google searches and found traditional-result clicks at 8% of visits when an AI summary appeared, versus 15% without one. Ahrefs later estimated that an AI Overview is associated with about 58% fewer clicks to the number-one organic result. A content site that used to be a small annuity is a harder business. That is a channel getting thinner. It is not a ban on charging for software.

AI made the other half of the complaint true. Building a thin tool is cheap, so lookalikes show up fast. Levels put the new price of execution at about $20 a month in a July 2026 post, and he described customers rebuilding a private copy instead of paying for a shallow product. That figure is his. The claim worth keeping is about a moat, not a price list. If the thing is easy to copy, the price people will pay for it falls. Money does not leave the category. It leaves the products that were only scarce because code was scarce.

Oversaturation is the same fact from the supply side. The crowded shelves are generic tools and products sold to other people trying to become indie hackers. A workflow buried in a job the founder actually knows is a different shelf.

Where money still shows up in public

The examples below are founder-reported, or Stripe-connected where noted. They are not a sample you can average.

Nomad List and Remote OK are the old proof that a small internet product can be a company. On 30 May 2019 Levels wrote that the two sites together were at about $83,333 a month. Four years, not four weeks. No funding round in that story.

Tony Dinh’s TypingMind is a later one. In November 2024 he wrote that it had crossed $1 million in the prior twelve months, about twenty months after launch. In October 2025 he put the product around $130,000–$160,000 a month. His books. Useful as a ceiling the feed stares at, useless as a median.

Marc Lou’s March 2026 recap is a portfolio, not a single app. He posted $78,120 for the month: TrustMRR $36,000, DataFast $21,000, ShipFast $9,200, CodeFast $8,300, plus smaller lines. He also wrote that AI had hurt the coding course and the boilerplate. The money moved inside the portfolio. ShipFast was still there. It was no longer the engine. His 2025 letter is a different mix again: about $1,032,000 including investing and payouts from X, not software alone.

Postiz is the clearest recent climb that is not only a screenshot. Nevo David has described a long grind into the teens of thousands of dollars in MRR, then a jump after the product became something AI agents could call. On 15 September 2026 TrustMRR showed about $222,851 MRR for Postiz, Stripe-connected on that page. The Postiz story keeps the founder-reported checkpoints and the verified print apart. Nobody published how much of that MRR came from agents. This page does not invent the split.

SuperX, Rob Hallam’s X tool, sat near $19,756 MRR on TrustMRR the morning of 25 September 2026, with his bio the same day still saying about $22,000 a month. Different meters. The Attention Playbook pre-sale is a course balance, not that MRR.

Bidding is also money, and it is a different kind. Jonathan Wilke’s week-one recap for outbid.lol, as written up by Steal What Works, reported about $221,000 in bid revenue. That is payment for a rank, not a retained customer. It can be real and still be a stunt economy. Treat it as the live plot on the timeline, not as evidence that the median SaaS got easier.

What still gets paid for

The products that keep showing up in the durable stories share a boring trait. Someone has a job the product sits inside, and the owner can reach that someone without renting the entire internet.

A narrow tool with a workflow is still a business when the clone is annoying to live in. Photo AI, TypingMind, a scheduler an agent can call, a jobs board, a directory people already check: different shapes, same test. The customer has a reason to stay that is not “the founder posted a milestone.”

An audience is an advantage, not a requirement. Lou has written that a following does not guarantee a startup will work. Levels’ nomad products predate the build-in-public gold rush. Postiz moved because a stranger’s article traveled, not because the founder began with a huge account. Distribution can be a niche you already have, a search page that still gets clicks, a partner, or a feed. “No audience” is not the same sentence as “no way to be found.”

Info products still sell, mostly to other makers. That revenue is real and it describes that audience. When the buyers are people trying to become you, the month can look healthy while the underlying tool market is soft. Lou’s own 2026 note about the course and the boilerplate is the caveat in the founder’s voice.

Ads and bidding sell attention directly. They pay when attention is the scarce thing and someone will buy a place in front of it. They are a worse model for a product that does not have attention yet. The split is on ads versus bidding.

The practical reading

You can still make money. You should not budget your life on the median screenshot.

The base rate from the verified-revenue feed is that a listed product is likely to make very little. The base rate from company-formation data is that more people start alone, and a first customer arrives faster than it did in 2020. The base rate from SaaS cohorts is that a million dollars of ARR inside three years stays uncommon. Hold those three at once.

If the plan is a generic wrapper around a model, the public argument in 2026 is that the customer may build it themselves. If the plan is a product inside a job you can name, with a price and a way to reach the buyer, the same year still has living companies. The starting conditions are still ordinary: something small, a charge, and a person who can hear about it. The dead part is the playbook that pretended those were easy.

FAQ